The difference in hit rate matters enormously and the existence of a hit tells you nothing about the denominator.
The size of the bet matters too - you could wipe out a hit with one bad pick if that bad pick is big enough!
If you don't have an estimate of the rate that you trust, you're just throwing money at dreams.
But it's WAY harder to get rich by being a pessimist than by being an optimist. And if you're a VC choosing where to invest primarily-other people's money, then you have no particular reason to try to talk them out of the hype.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:42:50.000Z","created_at_i":1791574970,"id":50025692,"options":[],"parent_id":50025484,"points":null,"story_id":50023450,"text":"The ex-OpenAI people who started Anthropic are doing pretty great right now.
VC is a hits business. Just one hit pays for 9 that didn't work out.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:25:48.000Z","created_at_i":1791573948,"id":50025484,"options":[],"parent_id":50024899,"points":null,"story_id":50023450,"text":"Same thing was said about Character AI, Cohere, etc. Companies started by the authors of the "Attention is all you need" paper. Look how that went... the team gets thrown around a lot as if it's the magic bullet. There is no magic bullet.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:37:43.000Z","created_at_i":1791571063,"id":50024899,"options":[],"parent_id":50024815,"points":null,"story_id":50023450,"text":"Probably their team. Ex-OpenAI people who have already proven they can ship and get buzz for what they're doing.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:32:10.000Z","created_at_i":1791570730,"id":50024815,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"What edge do they have over the market to justify such evaluation","title":null,"type":"comment","url":null},{"author":"prometheus1992","children":[{"author":"slopinthebag","children":[],"created_at":"2026-10-09T18:35:54.000Z","created_at_i":1791570954,"id":50024869,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"brb gonna wrap claude with a new form of prompting and raise 500 mil","title":null,"type":"comment","url":null},{"author":"Onavo","children":[],"created_at":"2026-10-09T18:37:19.000Z","created_at_i":1791571039,"id":50024894,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"You are used to dealing with companies where the money bags hold the power.
When you are in the middle of a boom cycle, it's the hottest company that has the advantage. Investing in them is a matter of privilege and they get to pick and choose.
Also, Canadian VCs are bottom of the barrel as far as VCs go.","title":null,"type":"comment","url":null},{"author":"besterman23","children":[{"author":"johnfn","children":[{"author":"besterman23","children":[],"created_at":"2026-10-09T18:57:36.000Z","created_at_i":1791572256,"id":50025150,"options":[],"parent_id":50024967,"points":null,"story_id":50023450,"text":"I\u2019m saying the bet is in them becoming THE System One Model company.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:42:24.000Z","created_at_i":1791571344,"id":50024967,"options":[],"parent_id":50024914,"points":null,"story_id":50023450,"text":"Isn't this the exact opposite? When people were saying that saying, the connotation was that IBM was an old, stodgy company that had been around forever. (These days I often think "No one got fired for choosing AWS"). Typesafe is a hot new startup that could, to my eyes, easily burst into flame or die in the next year.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:38:45.000Z","created_at_i":1791571125,"id":50024914,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"Probably the \u201cnobody ever got fired for buying IBM\u201d effect. If there\u2019s a use case for the tech, buying the most well known implementation of it will always be useful for people who want credit without the threat of blame. This funding is based entirely on the hype and a bet that TypeSafe will have name recognition.","title":null,"type":"comment","url":null},{"author":"kingcauchy","children":[],"created_at":"2026-10-09T18:40:47.000Z","created_at_i":1791571247,"id":50024941,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"It might be the people that are being acquired too, at huge inflated ai researchers salaries.
Acquired in the vc sense\u2026 not literal exit.","title":null,"type":"comment","url":null},{"author":"reticulates","children":[{"author":"nateb2022","children":[],"created_at":"2026-10-09T18:49:00.000Z","created_at_i":1791571740,"id":50025050,"options":[],"parent_id":50024947,"points":null,"story_id":50023450,"text":"I'd guess they justified the funding by revealing some grand scheme for a new product that they just need more runway to produce.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:40:56.000Z","created_at_i":1791571256,"id":50024947,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"The lack of a \u201cmoat\u201d is mostly irrelevant because success is not decided by who can or can\u2019t be cloned. TypeSafe invented[1] a new approach that became wildly popular almost immediately, if they can do that once, they can probably do it again. Venture capital is big bets, of course TypeSafe is going to fail, that\u2019s inevitable, but if it has even a 10% chance of capturing 1/10th the market cap of OpenAI then it is a great investment! Plus, money means nothing any more, they\u2019ve raised less at a lower valuation than Instinct, a personal assistant.
[1] not really but they did some innovative things and popularized a concept","title":null,"type":"comment","url":null},{"author":"phren0logy","children":[{"author":"JamesSwift","children":[{"author":"phren0logy","children":[],"created_at":"2026-10-10T23:42:40.000Z","created_at_i":1791675760,"id":50038259,"options":[],"parent_id":50028914,"points":null,"story_id":50023450,"text":"I have been running evals versus other models on my own data. For my use case, Jev comes out on top at least 2/3 of the time. This is purely about accuracy. For cost and speed, it is on top 100% of the time. I agree that people should test their own data against it before blindly assuming that it will get better outcomes.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T02:23:50.000Z","created_at_i":1791599030,"id":50028914,"options":[],"parent_id":50024952,"points":null,"story_id":50023450,"text":"I mean, a large majority of "pro jev" posts I see purely focus on the speed/cost and take their word on the accuracy. The entire product is "we guarantee the response will fit this schema, and will do it fast and cheap". Personally, I dont see how anything in this class can be "production ready" as a general classifier without the ability to fine tune on top.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:41:14.000Z","created_at_i":1791571274,"id":50024952,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"The API was duplicated, the results were not.","title":null,"type":"comment","url":null},{"author":"dvrp","children":[],"created_at":"2026-10-09T18:41:17.000Z","created_at_i":1791571277,"id":50024953,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"That is not how it works in the US.","title":null,"type":"comment","url":null},{"author":"InsideOutSanta","children":[],"created_at":"2026-10-09T18:49:22.000Z","created_at_i":1791571762,"id":50025054,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"> there is shit ton of due diligence at the tech level
Maybe in some cases. But counterexample, courtesy of The Information:
"It took just 15 minutes for Blue Owl executives to agree to invest up to $10 billion in future projects alongside real estate firm Primary Digital Infrastructure during their first in-person meeting two years ago, said Primary chief investment officer Bill Stein."
https://www.theinformation.com/articles/blue-owl-eyes-new-de...
AI seems to make some people lose their damned minds.","title":null,"type":"comment","url":null},{"author":"nlpnerd","children":[],"created_at":"2026-10-09T19:00:03.000Z","created_at_i":1791572403,"id":50025175,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"This is the difference between a "hot" company in a good ecosystem like SF. Yeah, the funds can take 2-3 months to do their due diligence. By the time it's done, the round has closed, and then what good is the due diligence?","title":null,"type":"comment","url":null},{"author":"bix6","children":[],"created_at":"2026-10-09T19:05:01.000Z","created_at_i":1791572701,"id":50025228,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"They got money and it needs to be put to work!","title":null,"type":"comment","url":null},{"author":"baobabKoodaa","children":[{"author":"hirako2000","children":[{"author":"baobabKoodaa","children":[{"author":"euleriancon","children":[],"created_at":"2026-10-09T20:22:09.000Z","created_at_i":1791577329,"id":50026119,"options":[],"parent_id":50025939,"points":null,"story_id":50023450,"text":"I think it is clear he is referring to zero shot classifiers with an LLM backbone. That tech has existed for a long time.","title":null,"type":"comment","url":null},{"author":"prometheus1992","children":[{"author":"baobabKoodaa","children":[{"author":"hbrn","children":[{"author":"baobabKoodaa","children":[{"author":"hbrn","children":[{"author":"baobabKoodaa","children":[{"author":"hbrn","children":[{"author":"baobabKoodaa","children":[],"created_at":"2026-10-09T23:15:24.000Z","created_at_i":1791587724,"id":50027772,"options":[],"parent_id":50027683,"points":null,"story_id":50023450,"text":"No, we're not at the point where you could ask Claude to do all of that, unless you have a super easy problem to begin with and/or you don't care about output quality. Feel free to link a counter example.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:05:08.000Z","created_at_i":1791587108,"id":50027683,"options":[],"parent_id":50027574,"points":null,"story_id":50023450,"text":"At this point there's no need to pay a data scientist. You can literally ask Claude to do everything for you: extract real examples, classify them, post-train a model, and ship an API.
Now, it could be viable if your business has literally hundreds of problems thats require classification. I just haven't seen those.
I treat the fact that almost noone was doing that as evidence that decision models aren't that useful/groundbreaking. That, and the fact that every single demo I saw was either fake (e.g. playing games), contrived, or plain wrong (e.g. using Jev for compaction).","title":null,"type":"comment","url":null},{"author":"majormajor","children":[{"author":"baobabKoodaa","children":[],"created_at":"2026-10-10T12:43:54.000Z","created_at_i":1791636234,"id":50032384,"options":[],"parent_id":50029078,"points":null,"story_id":50023450,"text":"Jev can do a subset of LLM things faster/cheaper. Jev is not better than big LLM's at anything.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T02:55:43.000Z","created_at_i":1791600943,"id":50029078,"options":[],"parent_id":50027574,"points":null,"story_id":50023450,"text":"Are there enough "tiny problems" for businesses in the world to justify this sort of money?
The appeal and claim of LLMs for businesses is solving big problems.
LLM valuations to solve tiny problems seems iffy.
--
Are there significant problem domains LLMs are bad at that Jev is good at? Vs just 'Jev can do a subset of LLM things faster/cheaper'?","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:52:26.000Z","created_at_i":1791586346,"id":50027574,"options":[],"parent_id":50027350,"points":null,"story_id":50023450,"text":"Yes, businesses are solving specific problems, but most businesses have more than 1 problem to solve. No, it is not economical to pay a data scientist to develop a custom model for each of your tiny problems. It is often much more economical to use a general purpose solution, like an LLM, or now, Jev.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:26:04.000Z","created_at_i":1791584764,"id":50027350,"options":[],"parent_id":50026826,"points":null,"story_id":50023450,"text":"> work on one specific problem, right? That is the opposite of "general purpose".
A business doesn't need Jev for the sake of Jev. Most business are solving specific problems.
And fine-tuning got a lot cheaper these days - I've seen claims here on HN that ~500 examples is enough to beat Jev.
> "nobody thought it's a good tradeoff", which again is false, there was huge demand for a cheap and accurate general purpose classifier
There wasn't. The hope is that there was a latent demand, but we've yet to see if it's truly latent or just manufactured.
Noone is saying "hell yeah, finally we got a general purpose classifier, my business needed it so much". The typical message is "this seems cool, let me see where I can apply it".
The fact that name itself is a play on Jevons Paradox illustrates that there was no demand until Jev was released.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:35:04.000Z","created_at_i":1791581704,"id":50026826,"options":[],"parent_id":50026678,"points":null,"story_id":50023450,"text":"When you say "trained classifiers", you are referring to models which are trained (or fine tuned) to work on one specific problem, right? That is the opposite of "general purpose".
Would Jev be more accurate in a specific task if it had been developed only for that task, as opposed to general purpose? Of course it would. So, sure, Jev is trading accuracy for generality. According to you "nobody thought it's a good tradeoff", which again is false, there was huge demand for a cheap and accurate general purpose classifier.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:19:06.000Z","created_at_i":1791580746,"id":50026678,"options":[],"parent_id":50026495,"points":null,"story_id":50023450,"text":"> Models we had before Jev had at most 2/3 of above qualities, but none of them were 3/3.
You're the one being deceptive here. Jev is trading accuracy, speed, and cost for generality. It's less accurate, slower and more expensive than trained classifiers. So it's still 2 out of 3, but with decimals. Maybe 2.2 out of 3 if I'm being charitable.
And the reason we didn't have that before is because nobody thought it's a good tradeoff.","title":null,"type":"comment","url":null},{"author":"prometheus1992","children":[{"author":"baobabKoodaa","children":[],"created_at":"2026-10-09T23:00:14.000Z","created_at_i":1791586814,"id":50027645,"options":[],"parent_id":50027239,"points":null,"story_id":50023450,"text":""Models on hf" (unspecified) are "FREE"? Like "free to download"? Sure, but nobody was talking about that. They cost money to run inference on. Unless you are talking about some tiny toy models that are useless for any non-toy problems. You clearly don't have any idea what you're talking about. Just stop, man.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:13:12.000Z","created_at_i":1791583992,"id":50027239,"options":[],"parent_id":50026495,"points":null,"story_id":50023450,"text":">>I was training custom ML models back in 2017
but you truly do sound like an angry 19 year old from your arguments.
- accurate - on what? on trust me bro benchmarks?
- zero-shot model are fundamentally general purpose.
- fast and cheap ; models on hf are FREE and fast enough.","title":null,"type":"comment","url":null},{"author":"gpugreg","children":[{"author":"baobabKoodaa","children":[{"author":"gpugreg","children":[{"author":"baobabKoodaa","children":[],"created_at":"2026-10-10T00:00:31.000Z","created_at_i":1791590431,"id":50028081,"options":[],"parent_id":50027883,"points":null,"story_id":50023450,"text":"Sounds like you are running a tiny toy model if you can get generations in under 5 ms? Typical response times from LLMs for typical "jev-like" queries from OpenAI and Anthropic are 2s-10s. Not milliseconds. Seconds. Same queries from Jev are like 0.2s. and the cost is 1000x.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:32:04.000Z","created_at_i":1791588724,"id":50027883,"options":[],"parent_id":50027714,"points":null,"story_id":50023450,"text":"llama.cpp is already extremely fast for single-token responses (<5 ms). I can't see Jev being faster when taking network latency into account, except maybe for multimodal inputs.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:09:36.000Z","created_at_i":1791587376,"id":50027714,"options":[],"parent_id":50027658,"points":null,"story_id":50023450,"text":"What you're missing is: cost and speed. Otherwise, it's very much like running an LLM and constraining the output.
I don't know how good Jev's "confidence scores" are, but I would be surprised if they were in any sense better than logits from some good LLM. One advantage of Jev here is that the confidence scores are easy to access. Most LLM API providers don't provide an easy/convenient way to access the logits. But that's a minor point, you could of course build something like this with LLMs (and many people have).","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:02:11.000Z","created_at_i":1791586931,"id":50027658,"options":[],"parent_id":50026495,"points":null,"story_id":50023450,"text":"
> I was training custom ML models back in 2017.\nJev is:
- accurate
- general purpose
- fast and cheap
Models we had before Jev had at most 2/3 of above qualities, but none of them were 3/3.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:48:03.000Z","created_at_i":1791578883,"id":50026397,"options":[],"parent_id":50025939,"points":null,"story_id":50023450,"text":"maybe you entered the AI space during the vibecoding era but there had been ton of useful models before that. especially zero shot models- both for text and images.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:04:30.000Z","created_at_i":1791576270,"id":50025939,"options":[],"parent_id":50025844,"points":null,"story_id":50023450,"text":"Unclear what you're referring to. Please stop making vague claims and be specific.","title":null,"type":"comment","url":null},{"author":"victorbjorklund","children":[],"created_at":"2026-10-10T12:09:05.000Z","created_at_i":1791634145,"id":50032103,"options":[],"parent_id":50025844,"points":null,"story_id":50023450,"text":"That is like saying ChatGPT wasn\u2019t anything new because text models had existed for years","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:56:08.000Z","created_at_i":1791575768,"id":50025844,"options":[],"parent_id":50025246,"points":null,"story_id":50023450,"text":"The tech behind it existed. They made it a specific product, and got replicated in days.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:06:09.000Z","created_at_i":1791572769,"id":50025246,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"> was already available
no, it was not
> was duplicated within a couple of days
was it already available or did it become available in a couple of days? it cant be both (neither is true, actually)","title":null,"type":"comment","url":null},{"author":"geoffschmidt","children":[{"author":"strgcmc","children":[],"created_at":"2026-10-10T14:53:29.000Z","created_at_i":1791644009,"id":50033578,"options":[],"parent_id":50025287,"points":null,"story_id":50023450,"text":"I think this is an astute comment and a reasonably accurate interpretation of how investors might be thinking about these opportunities.
At the same time, hopefully you would agree that this "self-reinforcing cycle" that you describe, if it really is the dominant way of thinking, is also a pretty clear sign that we're well into bubble territory, divorced from fundamentals.... Investing because you think your sheer act of investing is going to create a self-fulfilling prophecy of success while simultaneously having FOMO about someone else beating you to it? I dunno if I've seen a better description of what it looks like to make investment decisions from the POV of being deep inside a reality-distorted bubble while chasing mania.","title":null,"type":"comment","url":null},{"author":"prometheus1992","children":[],"created_at":"2026-10-10T15:04:08.000Z","created_at_i":1791644648,"id":50033673,"options":[],"parent_id":50025287,"points":null,"story_id":50023450,"text":"I think this is a reasonable take and I agree with you but its a bit unsettling that the event of fund raising has become a more important variable in the equation than the fundamental innovation itself. And I don't think this is a good long term strategy- it almost gives me "let's con our way out of this" vibe.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:09:30.000Z","created_at_i":1791572970,"id":50025287,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"There is a belief that there is going to be at least one more breakout success in startup AI labs - rather than OpenAI and Anthropic being the final word - and so investors want to own a part of whichever companies seem most likely to be that success. If you start from that premise and stack rank what company that might be, you could quite reasonably put TypeSafe toward the top of that list right now, based on the people at the company and the ability they've demonstrated to ship stuff that people care about and cut through the noise in a crowded space.
Also the situation isn't static. Investors know that the act of writing them a $870M check itself increases the chance that they'll be one of the winners, because that will attract more talent, customers, and funding to the company in a self-reinforcing cycle. And investors know that other investors know that, and that someone is going to write them that $870M check, so to some extent they're forced to think of the company as having already been successful at the fundraising and already having that momentum boost.
Only a small number of investors in the world can play the game at this level, because you have to smart enough to be right (often enough), and you have to be established enough to see the deals (be on every CEO's short list - because CEOs are only going to seriously pitch 5-10 VCs on a hot deal, if that). Otherwise you can't pull it off. Martin Casado and his team are among the few that can and I think their results reflect that.","title":null,"type":"comment","url":null},{"author":"dist-epoch","children":[{"author":"hirako2000","children":[],"created_at":"2026-10-09T20:04:44.000Z","created_at_i":1791576284,"id":50025942,"options":[],"parent_id":50025409,"points":null,"story_id":50023450,"text":"Whether a product has network effects makes a difference.
I would argue Dropbox did have a moat. It didn't merely store your data. It made it possible to make backup efficiently when bandwidth wasn't all that good.
Reading "no moat" so often is also tied to the fact those companies happen to be getting surreal valuations, at a quite early stage, showing no profit, building a tech that doesn't seem difficult to reproduce.","title":null,"type":"comment","url":null},{"author":"IshKebab","children":[],"created_at":"2026-10-09T20:17:56.000Z","created_at_i":1791577076,"id":50026070,"options":[],"parent_id":50025409,"points":null,"story_id":50023450,"text":"Dropbox had a super smooth UX that somehow nobody else replicated (seriously Google wtf). Instagram and Github won on network effects.","title":null,"type":"comment","url":null},{"author":"uuue","children":[],"created_at":"2026-10-09T22:40:07.000Z","created_at_i":1791585607,"id":50027484,"options":[],"parent_id":50025409,"points":null,"story_id":50023450,"text":"Instagram clearly did have one - Zuck had to acquire after internal efforts failed.","title":null,"type":"comment","url":null},{"author":"JamesSwift","children":[{"author":"dist-epoch","children":[{"author":"JamesSwift","children":[],"created_at":"2026-10-10T15:55:06.000Z","created_at_i":1791647706,"id":50034158,"options":[],"parent_id":50030891,"points":null,"story_id":50023450,"text":"Right, and yet the openai/anthropic models are still a tier above other options despite everyone gunning for them for at least one if not multiple years. While there are numerous jev-likes being spun up _in less than a month_ that are performing (as far as I can tell) up to and sometimes exceeding jev's accuracy.
This isnt the same as dropbox putting in a bunch of work to make the UX "just work", or even stripe doing the same for payments. This is "send me an api call that is in 1 of 3 forms, and get a response back in this form". The heavy lifting is in making that fast and cheap, and because they havent demonstrated a secret sauce in making it hard to outperform actual results that leaves a lot of room for competitors to get their existing models to run faster and cheaper for the same use case.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T08:32:24.000Z","created_at_i":1791621144,"id":50030891,"options":[],"parent_id":50028940,"points":null,"story_id":50023450,"text":"I meant more what was their moat when they launched, compared to their clones.
You can say the same thing about OpenAI/Anthropic, just an endpoint update, maybe a harness change if you use the CLI.
People have in fact been saying "what is the moat of OpenAI/Anthropic", yet here we are, trillion dollars valuation.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T02:28:29.000Z","created_at_i":1791599309,"id":50028940,"options":[],"parent_id":50025409,"points":null,"story_id":50023450,"text":"Dropbox: is I have to move my entire collection and setup the app on all my devices
Instagram: I have to move all my posts and also convince all my network to move over
Github: not a terrible example
If I decided to move off of jev tomorrow it would be an api key and a base api path update. Maybe 30 seconds of work.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:19:32.000Z","created_at_i":1791573572,"id":50025409,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"What was the moat of Dropbox? Of Instagram? Of GitHub? Or of countless other very successful startups when they started?
Anyone know when this "have no moat" meme appeared? Even 5 years ago I don't remember seeing it on every post.","title":null,"type":"comment","url":null},{"author":"cmrdporcupine","children":[],"created_at":"2026-10-09T19:43:45.000Z","created_at_i":1791575025,"id":50025703,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"Yeah your problem and my problem and others around here is the word you just said there... "Toronto." Canadian investors are risk averse as hell. And cheap. They can make more money helping sell bitumen or real estate, why bother with arcane tech?
And if you could put the words "Bay Area" or "Stanford" or "San Francisco" next to your name... different story.
The VCs are not buying the idea or the tech, they're investing in the people. And they invest in a formula that has already worked for them before to make big coin. Prop somebody up, let them hire like crazy, and then get them get acquired, and then cash out. They don't care if it fails if they can make it succeed 1/200 times.
Canadian investors want you to have already succeeded before they help you succeed a tiny bit more.","title":null,"type":"comment","url":null},{"author":"fidotron","children":[{"author":"redanddead","children":[],"created_at":"2026-10-09T21:13:26.000Z","created_at_i":1791580406,"id":50026625,"options":[],"parent_id":50025741,"points":null,"story_id":50023450,"text":"Canada doesn\u2019t have throwing around money like in the US. We have resource extraction -> export money that\u2019s it","title":null,"type":"comment","url":null},{"author":"cmrdporcupine","children":[],"created_at":"2026-10-09T21:32:49.000Z","created_at_i":1791581569,"id":50026806,"options":[],"parent_id":50025741,"points":null,"story_id":50023450,"text":"There are boatloads of tax breaks and incentives that mitigate all the financial stuff and make running a startup here just fine honestly.
But that does nothing to make up for the terrible investment community. Getting started here requires already being started.
When I briefly worked for a Toronto startup, it was like all of them went to the same private boy's schools together as kids. It was a status club.
I jumped ship to an American startup and made almost double the money dealt with 0% of the bullshit and they were bought by Google the next year.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:48:32.000Z","created_at_i":1791575312,"id":50025741,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"> sat in fund raising meetings with VCs in toronto
There's your problem. The single biggest thing every Canadian VC is trying to figure out is "why are these people asking us for money when if they were any good they'd be in the US" so by simply asking them you're already signalling something bad. A lot of their enthusiasm for process is based on this suspicion and also that the entire industry is just a way for various professional services to extract most of the investment money, since that's the game they're so used to playing with the government.
There are some Canadian VCs earnestly trying to improve but they are overwhelmingly hilariously conservative and focused on unimportant signals over reality. This is one (but not all) of the major factors that drive basically every remotely ambitious Canadian company to run a corp in Delaware and go for funding from the US. The tax situation is the other major contributor.","title":null,"type":"comment","url":null},{"author":"havercosine","children":[{"author":"majormajor","children":[],"created_at":"2026-10-10T02:53:05.000Z","created_at_i":1791600785,"id":50029064,"options":[],"parent_id":50025831,"points":null,"story_id":50023450,"text":"> Its not normal time in SF/Bay Area. For better or worse, VCs in this city/region are thinking very differently on AI bets.
"Throw a bunch of money at copycats in the trend of the day" is very normal in SF/SV for the last several decades, going back to the dotcom boom.
The definition of "a bunch" has changed but so many "crypto for X" or "recommendations for X" or "uber for X" or "social for X", etc, things raised amounts that seemed wildly divorced from their market position.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:55:12.000Z","created_at_i":1791575712,"id":50025831,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"Its not normal time in SF/Bay Area. For better or worse, VCs in this city/region are thinking very differently on AI bets.
I think the key differentiator was that a team found a whitespace in what ChatGPT was doing, main comes from the same pedigree and team is as conscious of marketing as their product. SF VCs love these out of the box challengers, and people are claiming to replicate doesn't seem to matter.
The amount raised feels surprising but again entire SF/US AI scene is primarily "add moar layers and GPU" one trick ponies at this point.","title":null,"type":"comment","url":null},{"author":"HarHarVeryFunny","children":[],"created_at":"2026-10-09T19:57:08.000Z","created_at_i":1791575828,"id":50025862,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"A very major part of Jev is the cost and speed. Yes, classification is/will be a commodity business, just like LLMs are, and similarly there is no moat only production cost and pricing.
Yes, anyone can wrap a decisions API around an LLM, but so what? If you want to compete then you need to compete on price, and it's not clear if OpenAI and/or Anthropic are able or willing to do that without building a custom architecture, and even then is a race to the bottom on pricing really what they want to pursue?
I'm not sure if OpenAI have announced pricing for their Decisions API, but they have said it's based on Luna which costs $0.10/M input, not even remotely competitive with Jev's $0.04/M input, which I'd expect has some headroom built into it.
Assuming that the architecture behind Jev is not just an LLM, and gives them some inherent efficiency/cost and speed advantage, then the question is whether OpenAI and Anthropic really want to duplicate this and have a race to the bottom on pricing for what may be a large part of the business automation market they are addressing. Is that what they want as their IPO pitch - we're selling potatoes, and think can grow them cheaper than Typesafe ?","title":null,"type":"comment","url":null},{"author":"binlog","children":[],"created_at":"2026-10-09T21:06:46.000Z","created_at_i":1791580006,"id":50026552,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"Because every VC knows that one of OpenAI/Anthropic/Nvidia/Microsoft/Google/Meta/SpaceXAI/AMD/Stripe... will acquire them within the next year for talent alone.","title":null,"type":"comment","url":null},{"author":"smrtinsert","children":[],"created_at":"2026-10-09T21:22:48.000Z","created_at_i":1791580968,"id":50026717,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"When you frame it as if we're in the Pets.com era of AI the continued gold rush makes sense.","title":null,"type":"comment","url":null},{"author":"baby","children":[],"created_at":"2026-10-10T12:40:40.000Z","created_at_i":1791636040,"id":50032366,"options":[],"parent_id":50024839,"points":null,"story_id":50023450,"text":"My guess is users, they must have had a MASSIVE user growth since they have launched to justify this","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:33:33.000Z","created_at_i":1791570813,"id":50024839,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"I really don't understand how this can be. I have sat in fund raising meetings with VCs in toronto and my experience is that there is shit ton of due diligence at the tech level. a product which has no moat, was already available, was duplicated within a couple of days is valued at 7B - i thought we were past the peak of the hype cycle.","title":null,"type":"comment","url":null},{"author":"jghn","children":[],"created_at":"2026-10-09T18:39:56.000Z","created_at_i":1791571196,"id":50024930,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Every time I see these headlines I wonder why the Scala company is back in the news","title":null,"type":"comment","url":null},{"author":"bel8","children":[],"created_at":"2026-10-09T18:40:18.000Z","created_at_i":1791571218,"id":50024936,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Do they have patents over jev related tech or something valuable to justify this?","title":null,"type":"comment","url":null},{"author":"armcat","children":[{"author":"dkersten","children":[{"author":"devin","children":[{"author":"baobabKoodaa","children":[{"author":"homarp","children":[{"author":"baobabKoodaa","children":[{"author":"devin","children":[{"author":"baobabKoodaa","children":[{"author":"rpdillon","children":[{"author":"baobabKoodaa","children":[{"author":"devin","children":[{"author":"baobabKoodaa","children":[{"author":"devin","children":[{"author":"baobabKoodaa","children":[],"created_at":"2026-10-11T14:56:31.000Z","created_at_i":1791730591,"id":50043983,"options":[],"parent_id":50034624,"points":null,"story_id":50023450,"text":"No, what you said was:
> What it does was published in arxiv back in 2025
This is false. You made a statement which was false.
If you want to say that the author of the paper made a really bad knockoff of Jev in 2026 after Jev was published, then say that.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T16:43:25.000Z","created_at_i":1791650605,"id":50034624,"options":[],"parent_id":50030496,"points":null,"story_id":50023450,"text":"I said the author of the paper created a generalized version because they posted about doing so on reddit. They way they described the work, it sounded trivial.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T07:14:58.000Z","created_at_i":1791616498,"id":50030496,"options":[],"parent_id":50029576,"points":null,"story_id":50023450,"text":"The comment I was responding to describes Jev's significance as being specifically about embeddings:
> Jev is basically the embeddings side of an LLM
This is how the paper you are referencing is describing the "key contribution" as it relates to embeddings:
> Novel state representation techniques using Azure OpenAI embeddings (3072 dimensions) with sales-specific features
And now you are claiming that the author of the paper created:
> "generalized" version of same
It's a bit hard to guess what you are trying to say, but if I were to steelman your argument, I would guess that you mean: training a model with a vocabulary that has some tokens representing things like "OPTION_A", "OPTION_B" is in your mind the same thing as "generalized version of sales-specific features in embeddings"? Is this what you were trying to say?","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T04:31:34.000Z","created_at_i":1791606694,"id":50029576,"options":[],"parent_id":50027550,"points":null,"story_id":50023450,"text":"Hard to take you seriously, fam. The distance between these two things is trivial. The author of the paper implemented the "generalized" version of same in less than 24 hours. It's not novel, it is a continuation of what already existed in a boring way","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:48:38.000Z","created_at_i":1791586118,"id":50027550,"options":[],"parent_id":50026889,"points":null,"story_id":50023450,"text":"No. It can't simultaneously be both general purpose and having task-specific embeddings.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:38:39.000Z","created_at_i":1791581919,"id":50026889,"options":[],"parent_id":50025772,"points":null,"story_id":50023450,"text":"> Novel state representation techniques using Azure OpenAI embeddings (3072 dimensions) with sales-specific features
Jev is basically the embeddings side of an LLM. Yes, it's a good idea, but the moat is non-existent.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:50:33.000Z","created_at_i":1791575433,"id":50025772,"options":[],"parent_id":50025631,"points":null,"story_id":50023450,"text":"Bullshit. Below is copypaste from the paper in the section that outlines the "key contributions" of the paper. As you can see, it is focused on one specific problem: predicting sales conversions. So if you were to take this system and use it for some other task ("evaluate customer mood" for example), it would not work. Because, again, it is not describing a general purpose solution. It is describing a solution that is specific to one problem: sales conversions.
Copypasta:
\u2022 A reinforcement learning architecture specifically designed for sales conversation analysis and conversion\nprediction
\u2022 A synthetic data generation pipeline leveraging GPT-4O to create diverse and realistic sales conversations
\u2022 Novel state representation techniques using Azure OpenAI embeddings (3072 dimensions) with sales-specific\nfeatures
\u2022 A meta-learning approach enabling the system to express confidence in its predictions based on conversation similarity to training data
\u2022 Integration mechanisms providing real-time guidance within existing sales platforms
\u2022 Extensive comparative evaluation demonstrating significant performance improvements over LLM-based approaches","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:37:37.000Z","created_at_i":1791574657,"id":50025631,"options":[],"parent_id":50025482,"points":null,"story_id":50023450,"text":"What do you mean? Jev is trivially different from what is described in this paper.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:25:38.000Z","created_at_i":1791573938,"id":50025482,"options":[],"parent_id":50025453,"points":null,"story_id":50023450,"text":"Yeah, I figured it was gonna be this one.
"SalesRLAgent: A Reinforcement Learning Approach for Real-Time Sales Conversion Prediction and Optimization"
Jev is a general-purpose thing. That is a specific-purpose thing. General-purpose thing is not the same as specific-purpose thing. What makes people think these are the same thing? I don't get it.","title":null,"type":"comment","url":null},{"author":"zwaps","children":[],"created_at":"2026-10-09T20:10:18.000Z","created_at_i":1791576618,"id":50025998,"options":[],"parent_id":50025453,"points":null,"story_id":50023450,"text":"This is a fine-tuned model. The author even states that the model is competitive with Jev only if fine-tuned on the evaluation at hand.
Literally misses the point of Jev, which you don't need to fine-tune to get accuracy nor - and no other model has this - some sort of out of sample calibration","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:22:33.000Z","created_at_i":1791573753,"id":50025453,"options":[],"parent_id":50025399,"points":null,"story_id":50023450,"text":"https://arxiv.org/abs/2503.23303
and https://arxiv.org/abs/2510.01237","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:18:37.000Z","created_at_i":1791573517,"id":50025399,"options":[],"parent_id":50025159,"points":null,"story_id":50023450,"text":"What specific arxiv paper are you referencing here?","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:58:40.000Z","created_at_i":1791572320,"id":50025159,"options":[],"parent_id":50025032,"points":null,"story_id":50023450,"text":"Jev did not take years to develop. What it does was published in arxiv back in 2025. TypeSafe just marketed it.","title":null,"type":"comment","url":null},{"author":"hbrn","children":[{"author":"dkersten","children":[{"author":"hbrn","children":[],"created_at":"2026-10-09T20:51:20.000Z","created_at_i":1791579080,"id":50026429,"options":[],"parent_id":50026150,"points":null,"story_id":50023450,"text":"> zero hallucination
Plenty has been said about this claim. If you're still falling for this, I feel sorry for you.
If you remove wheels from your car, your car will get a "no speeding ticket" property, and yet there's nothing exciting about it.
> You can\u2019t get that with a fine tuned LLM
Of course you can. All these claims are nothing but marketing.","title":null,"type":"comment","url":null},{"author":"user43928","children":[{"author":"dkersten","children":[],"created_at":"2026-10-10T07:57:00.000Z","created_at_i":1791619020,"id":50030726,"options":[],"parent_id":50026961,"points":null,"story_id":50023450,"text":"I\u2019m not claiming you can\u2019t get some of the properties in some ways with LLMs, just that Jev is the sum of its parts, not just one or two properties.
LLMs can produce structured output by limiting the next token based on a grammar, that ensures correctly formed output.
But (presumably, again I haven\u2019t seen Jev\u2019s insides) Jev doesn\u2019t need to do that, it just has to output probabilities for each answer, and can do it natively without artificially limiting output tokens.
So jev is cheap, fast, never produces malformed output, its output always carries correct semantic meaning (but this doesn\u2019t mean it always answers correctly), and it separates correct from prompt (which if it truly does that is the most exciting part).","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:44:59.000Z","created_at_i":1791582299,"id":50026961,"options":[],"parent_id":50026150,"points":null,"story_id":50023450,"text":"I understand most major model providers support passing a JSON schema that is strictly followed in the output, accomplishing the same 'zero hallucination' and prompt injection resistance.
The only difference I am aware of is that probabilities are better calibrated with these decision models compared to regular LLMs which can output hallucinated numbers where your schema allows a number.","title":null,"type":"comment","url":null},{"author":"PunchyHamster","children":[{"author":"dkersten","children":[],"created_at":"2026-10-12T09:42:27.000Z","created_at_i":1791798147,"id":50053805,"options":[],"parent_id":50031078,"points":null,"story_id":50023450,"text":"Can you share a sample prompt demonstrating it? I\u2019d be interested to see, especially since you\u2019re saying it\u2019s incredibly easy (20 seconds would mean it\u2019s easier than on most normal LLMs even).","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T09:13:36.000Z","created_at_i":1791623616,"id":50031078,"options":[],"parent_id":50026150,"points":null,"story_id":50023450,"text":"Took me about 20 seconds to prompt inject the demo instance.... y'all would be great scammer targets","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:25:40.000Z","created_at_i":1791577540,"id":50026150,"options":[],"parent_id":50025632,"points":null,"story_id":50023450,"text":"My point is that it\u2019s not replicated. You replicate the accuracy, but not the other properties. The Jev-competitors only proved that you can get or beat the accuracy, nothing about the other properties. Especially the \u201czero hallucination\u201d output and the (if it works how the documentation make it sound) prompt injection resistant architecture. You can\u2019t get that with a fine tuned LLM.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:37:49.000Z","created_at_i":1791574669,"id":50025632,"options":[],"parent_id":50025032,"points":null,"story_id":50023450,"text":"> I strongly doubt that a copycat that was put together within days after Jev\u2019s release will be able to match it on a sun of its properties
But why? If the simplest way to achieve Jev's capabilities (accuracy, cost, latency) is by fine tuning a small model, what makes you think that this isn't exactly what Typesafe did?
And even if they did something different - what makes you think it was a good idea in the first place, given how easy their results were replicated without any "secret sauce"?","title":null,"type":"comment","url":null},{"author":"TeMPOraL","children":[{"author":"mrinterweb","children":[{"author":"lifeisloving","children":[],"created_at":"2026-10-09T23:19:08.000Z","created_at_i":1791587948,"id":50027792,"options":[],"parent_id":50026785,"points":null,"story_id":50023450,"text":"There will always only be a very small percentage of people who want to discover and build things, even with llms, its a very small subset of people though, and most people want off the shelf solutions.
Also Jevs purpose isnt to become its own thing. It will get aquired in 18 months by one of Andressen Horowitz's incestuous circle of companies and everyone will make money, and the person who buys it wont necessarily care if Jev itself makes them a ton of money. They're just passing chips around the table.","title":null,"type":"comment","url":null},{"author":"TeMPOraL","children":[],"created_at":"2026-10-10T08:20:26.000Z","created_at_i":1791620426,"id":50030835,"options":[],"parent_id":50026785,"points":null,"story_id":50023450,"text":"Jev is the known idea, but with a slick website with grandiose claims and an impressive Doom demo on. It's also a catchy name - it caught on instantly, but mostly as a shorthand: it's easier to say "Jev" and "like Jev" than "classifier models" (or rather "<descriptive explanation of a specific shape of> classifier models").","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:30:19.000Z","created_at_i":1791581419,"id":50026785,"options":[],"parent_id":50026165,"points":null,"story_id":50023450,"text":"That probably explains why there were so many competitors around withing days of the Jev announcement. They are not starting with a moat, and there doesn't seem to be any moat in sight. Just buzzword recognition because everything is comparing to "jev".","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:27:05.000Z","created_at_i":1791577625,"id":50026165,"options":[],"parent_id":50025032,"points":null,"story_id":50023450,"text":"Jev is something your favorite LLM could zero-shot months ago, if you pointed it to the right arXiv paper (some of which are linked in this thread).","title":null,"type":"comment","url":null},{"author":"creatonez","children":[{"author":"dkersten","children":[],"created_at":"2026-10-10T08:00:54.000Z","created_at_i":1791619254,"id":50030743,"options":[],"parent_id":50029939,"points":null,"story_id":50023450,"text":"It separates context/state from prompt/questions. That\u2019s a key ingredient for prompt injection hardening. If it is designed and trained to never treat state as instruction, and you only ever put untrusted user input into state, and never the questions, then why wouldn\u2019t it be?
The jev documentation says this is what you should do. It doesn\u2019t mean that it\u2019s actually designed for prompt injection resistance, but it could be. We won\u2019t know for sure until they either release more details or someone proves otherwise.
But the split is one that doesn\u2019t exist in normal LLMs and it\u2019s the exact split that is needed for immunity or resistance to prompt injection.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T05:40:43.000Z","created_at_i":1791610843,"id":50029939,"options":[],"parent_id":50025032,"points":null,"story_id":50023450,"text":"Jev is almost certainly not immune to prompt injection","title":null,"type":"comment","url":null},{"author":"nowittyusername","children":[],"created_at":"2026-10-10T15:13:27.000Z","created_at_i":1791645207,"id":50033757,"options":[],"parent_id":50025032,"points":null,"story_id":50023450,"text":"JEV's moat is calibration nothing else. All other llms excel at system one like judgement versus JEV, but that's not a dunk on JEV. Calibration is very important for some niche use cases.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:47:35.000Z","created_at_i":1791571655,"id":50025032,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Most of them appear to be small LLM\u2019s fine tuned for the role.
That\u2019s a different set of properties in terms of size, cost, and latency. Jev (apparently, not like I\u2019ve seen its insides) is extremely cheap, extremely fast, doesn\u2019t cost any output tokens as it speaks the output natively, can\u2019t get the output wrong because it speaks the format natively, and (presumably based on the docs), the context is separate from the question, meaning it should be immune (or at least highly resistant) to prompt injection attacks.
It\u2019s not just about the accuracy of the result, it\u2019s a collection of all the properties that make Jev interesting.
Jev took years to develop, I strongly doubt that a copycat that was put together within days after Jev\u2019s release will be able to match it on a sun of its properties. Even if fine tuned LLMs can outperform it on raw accuracy.","title":null,"type":"comment","url":null},{"author":"mlmonkey","children":[{"author":"bayesianbot","children":[],"created_at":"2026-10-09T19:04:33.000Z","created_at_i":1791572673,"id":50025222,"options":[],"parent_id":50025079,"points":null,"story_id":50023450,"text":"That is their Python SDK version, not Python version","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:51:41.000Z","created_at_i":1791571901,"id":50025079,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"OpenAI's "Decisions" library has this in requirements:
To run the SDK examples below, use these OpenAI SDK versions or later: Python 3.26.0,
I thought Pythin 3.15.0 just came out, 3.26.0 must be really far off?","title":null,"type":"comment","url":null},{"author":"scottyah","children":[{"author":"rockwotj","children":[],"created_at":"2026-10-10T02:02:41.000Z","created_at_i":1791597761,"id":50028778,"options":[],"parent_id":50025104,"points":null,"story_id":50023450,"text":"I have ran a bunch of evals on production use cases where we are using Jev and OpenAI's model isn't close, but I'm hopeful about some of the OSS ones. IDK it's very cheap so I don't really feel like I need to look for an alternative. Why root for OpenAI over Typesafe? Isn't more diversity in the market good?","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:53:41.000Z","created_at_i":1791572021,"id":50025104,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"> OpenAI's own Decisions API [1] beats it
Have you heard that from a different source than OpenAI? From what I'd heard other models haven't gotten close, and the open source ones are like running gemma4 E2B against Opus 5.5- sure, the API calls go in and are returned the same but the quality isn't close.","title":null,"type":"comment","url":null},{"author":"nlpnerd","children":[],"created_at":"2026-10-09T18:58:06.000Z","created_at_i":1791572286,"id":50025155,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"You are assuming that the VCs have done their due diligence. For a "hot" company like Typesafe AI, most likely little due diligence was done. That's the way it's played.","title":null,"type":"comment","url":null},{"author":"baobabKoodaa","children":[{"author":"ricericerice","children":[{"author":"baobabKoodaa","children":[],"created_at":"2026-10-09T19:34:52.000Z","created_at_i":1791574492,"id":50025602,"options":[],"parent_id":50025584,"points":null,"story_id":50023450,"text":"And even if you do have that, you haven't made your own Jev, because Jev is a general-purpose thing, whereas what you have built is a specific-purpose thing.","title":null,"type":"comment","url":null},{"author":"9dev","children":[],"created_at":"2026-10-10T07:45:01.000Z","created_at_i":1791618301,"id":50030663,"options":[],"parent_id":50025584,"points":null,"story_id":50023450,"text":"\u2026and a team with experience and the hardware and the workflows to do this at scale, repeatedly to avoid drift, with proper infrastructure for evaluation and testing, and all of this work is somehow the core domain of your business.
Then yes, easy!","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:33:36.000Z","created_at_i":1791574416,"id":50025584,"options":[],"parent_id":50025216,"points":null,"story_id":50023450,"text":"you can easily finetune your own*
*if you have a sufficiently sized and quality dataset for the specific classifications you're targeting","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:04:00.000Z","created_at_i":1791572640,"id":50025216,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"> you can easily finetune your own
no, you can't, and it's unclear why you would think this.","title":null,"type":"comment","url":null},{"author":"ttul","children":[{"author":"Oras","children":[{"author":"brink","children":[],"created_at":"2026-10-09T19:33:06.000Z","created_at_i":1791574386,"id":50025581,"options":[],"parent_id":50025481,"points":null,"story_id":50023450,"text":"Either the investors know something we don't, or the market is irrational.","title":null,"type":"comment","url":null},{"author":"rsalus","children":[{"author":"SaltyBackendGuy","children":[{"author":"Amekedl","children":[],"created_at":"2026-10-10T12:05:38.000Z","created_at_i":1791633938,"id":50032076,"options":[],"parent_id":50027860,"points":null,"story_id":50023450,"text":"Always has been indeed.\nMarketing is the closest thing to a real superpower if done right, no matter how successful your product in practice is, or if it even does anything at all (see any and all youtube sponsorships kind of...)","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:29:03.000Z","created_at_i":1791588543,"id":50027860,"options":[],"parent_id":50025591,"points":null,"story_id":50023450,"text":"Always has been imo.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:33:55.000Z","created_at_i":1791574435,"id":50025591,"options":[],"parent_id":50025481,"points":null,"story_id":50023450,"text":"I feel like half the game now is marketing though, so I can see why they'd be attractive to an investor. Maybe if they scale they can come up with something.","title":null,"type":"comment","url":null},{"author":"ModernMech","children":[],"created_at":"2026-10-09T19:34:38.000Z","created_at_i":1791574478,"id":50025600,"options":[],"parent_id":50025481,"points":null,"story_id":50023450,"text":"It\u2019s the classic SV flip. You scale your investors, your executive team, your sales people, hire a bunch of engineering you don\u2019t need, then sell the company. The company\u2019s product doesn\u2019t matter, the company is the product.","title":null,"type":"comment","url":null},{"author":"mococa","children":[{"author":"shdh","children":[],"created_at":"2026-10-09T20:38:28.000Z","created_at_i":1791578308,"id":50026289,"options":[],"parent_id":50025801,"points":null,"story_id":50023450,"text":"So did Oracle with relational databases by that logic","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:52:52.000Z","created_at_i":1791575572,"id":50025801,"options":[],"parent_id":50025481,"points":null,"story_id":50023450,"text":"> Their 2 years of stealth was replicated in 2 weeks.
Actually they stolen the idea from a paper.","title":null,"type":"comment","url":null},{"author":"tiborsaas","children":[{"author":"throwaway7783","children":[{"author":"PunchyHamster","children":[{"author":"throwaway7783","children":[],"created_at":"2026-10-11T20:38:09.000Z","created_at_i":1791751089,"id":50047586,"options":[],"parent_id":50031057,"points":null,"story_id":50023450,"text":"Marketing always was a forcing function. The barrier to product building has dropped immensely now, so it is more about "finesse" part of the product in that area.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T09:08:37.000Z","created_at_i":1791623317,"id":50031057,"options":[],"parent_id":50026388,"points":null,"story_id":50023450,"text":""now"? It always was. Pre-AI we still had ton of great apps "losing" just because the couldn't be brought before the eyes of people that would like them","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:47:17.000Z","created_at_i":1791578837,"id":50026388,"options":[],"parent_id":50026358,"points":null,"story_id":50023450,"text":"100%. Product finesse + Marketing is now the "moat".","title":null,"type":"comment","url":null},{"author":"tyre","children":[{"author":"wordpad","children":[{"author":"Bishonen88","children":[],"created_at":"2026-10-10T12:14:50.000Z","created_at_i":1791634490,"id":50032157,"options":[],"parent_id":50028199,"points":null,"story_id":50023450,"text":"Kiro came out after both Claude and codex, so not sure that any enterprise stuck to kiro unless you mean that they stick to aws in general. I'd love to hear about usage of kiro even in vague terms compared to the competitors. Naively thinking, I'd guess it's 1% or less of Claude code.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T00:19:23.000Z","created_at_i":1791591563,"id":50028199,"options":[],"parent_id":50027534,"points":null,"story_id":50023450,"text":"Enterprises are very sticky and they arent even chosing between claude and codex, many are looking at something like Kiro.","title":null,"type":"comment","url":null},{"author":"tiborsaas","children":[],"created_at":"2026-10-10T12:00:27.000Z","created_at_i":1791633627,"id":50032046,"options":[],"parent_id":50027534,"points":null,"story_id":50023450,"text":"I think it's a ridiculous amount of money they got, but we don't know what's in the pitch deck. Investors are also blind to most of these things and take the pedigree of the founders above else.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:46:50.000Z","created_at_i":1791586010,"id":50027534,"options":[],"parent_id":50026358,"points":null,"story_id":50023450,"text":"Yeah but there are also network effects of using fitness apps. Every engineer I know switches easily between Codex and Claude Code when one gets better than the other.
Jev has been around for a couple weeks. Cost and performance matter more than anything. Staying with an existing provider (the # of people choosing Jev without already having a frontier API key is probably zero?) is way easier than this.
What the hell are we even talking about.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:44:09.000Z","created_at_i":1791578649,"id":50026358,"options":[],"parent_id":50025481,"points":null,"story_id":50023450,"text":"You can replicate any fitness app in no time and you will make close to $0. Brand recognition matters a lot.","title":null,"type":"comment","url":null},{"author":"clickety_clack","children":[{"author":"jeremyjh","children":[],"created_at":"2026-10-09T21:40:49.000Z","created_at_i":1791582049,"id":50026923,"options":[],"parent_id":50026728,"points":null,"story_id":50023450,"text":"You say this, while Hermes Agent has been at top of the openrouter.ai leaderboard for several months and currently has 3X the token usage of OpenClaw.","title":null,"type":"comment","url":null},{"author":"real0mar","children":[{"author":"Espressosaurus","children":[],"created_at":"2026-10-09T22:43:26.000Z","created_at_i":1791585806,"id":50027511,"options":[],"parent_id":50027285,"points":null,"story_id":50023450,"text":"And Muse is what normal people use instead","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:17:39.000Z","created_at_i":1791584259,"id":50027285,"options":[],"parent_id":50026728,"points":null,"story_id":50023450,"text":"What? OpenClaw has been completely replaced by Hermes and others in the discourse","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:24:04.000Z","created_at_i":1791581044,"id":50026728,"options":[],"parent_id":50025481,"points":null,"story_id":50023450,"text":"It\u2019s like openclaw. There was a bunch of technically better ones that came along afterwards, but nobody remembers what any of them were called.","title":null,"type":"comment","url":null},{"author":"OtherShrezzing","children":[{"author":"vasco","children":[],"created_at":"2026-10-10T07:40:33.000Z","created_at_i":1791618033,"id":50030631,"options":[],"parent_id":50030235,"points":null,"story_id":50023450,"text":"> People who can find things of value, and execute on those things, are worth investing in
Investing isn't binary. You can be investible but not at a 7.5b valuation.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T06:27:43.000Z","created_at_i":1791613663,"id":50030235,"options":[],"parent_id":50025481,"points":null,"story_id":50023450,"text":"Their 2 years of stealth involved a lot more than building their model.
People who can find things of value, and execute on those things, are worth investing in - even if that thing of value is replicated in short order.
Nobody was looking at decision models, but now that Jev has surfaced, everybody is.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:25:28.000Z","created_at_i":1791573928,"id":50025481,"options":[],"parent_id":50025244,"points":null,"story_id":50023450,"text":"Rapidly? Their 2 years of stealth was replicated in 2 weeks.
I give it to them for creating the hype (good marketing), and for making a useful classifier. Not sure what they would scale rapidly though.","title":null,"type":"comment","url":null},{"author":"sailfast","children":[],"created_at":"2026-10-10T13:18:09.000Z","created_at_i":1791638289,"id":50032648,"options":[],"parent_id":50025244,"points":null,"story_id":50023450,"text":"Is anybody really looking to scale a company\u2019s people rapidly these days? It\u2019s just not necessary.
Most of the money will be burned in data centers as they struggle to sell tokens at a loss.","title":null,"type":"comment","url":null},{"author":"make3","children":[],"created_at":"2026-10-10T20:44:26.000Z","created_at_i":1791665066,"id":50036986,"options":[],"parent_id":50025244,"points":null,"story_id":50023450,"text":"Jev and decision models are *a derivative* of a pretrained LLM trunk, and one that's super easy to build once you have an LLM. Put a single step fully connected NN on the transformer, create a training dataset for it, add fancy but bread and butter multi step decision RL, nothing crazy.
*They have zero moat* against established AI companies, as the recent copies show, and will do worse than frontier labs.*","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:06:02.000Z","created_at_i":1791572762,"id":50025244,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"But Jev established the branding and investors are betting that Jev will be acquired by one of the big labs soon - and if they aren't, the money itself can create a positive outcome by allowing Jev to hire incredible talent and scale the company rapidly.","title":null,"type":"comment","url":null},{"author":"doctorpangloss","children":[],"created_at":"2026-10-09T19:18:29.000Z","created_at_i":1791573509,"id":50025397,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":""It doesn't matter"
By all means, become an A16Z LP.","title":null,"type":"comment","url":null},{"author":"amelius","children":[],"created_at":"2026-10-09T19:29:11.000Z","created_at_i":1791574151,"id":50025533,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"I mean I'm already ditching my Apple stocks because soon AI will be able to replicate iOS and MacOS.","title":null,"type":"comment","url":null},{"author":"827a","children":[{"author":"nlpnerd","children":[{"author":"TeMPOraL","children":[{"author":"nlpnerd","children":[],"created_at":"2026-10-10T04:56:41.000Z","created_at_i":1791608201,"id":50029700,"options":[],"parent_id":50026129,"points":null,"story_id":50023450,"text":"One may consider it "rigidity" but many application backends (including LLM powered ones) never needed the full unconstrained capabilities of an LLM. What people needed was the zero-shot capabilities over tasks as opposed to have to train a classifier, span model, etc for every task.
Constraints can be an advantage in many systems. Consider typed and untyped programming languages. Many advantages with typed languages over the latter in terms of development and efficiency.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:22:55.000Z","created_at_i":1791577375,"id":50026129,"options":[],"parent_id":50025702,"points":null,"story_id":50023450,"text":"Skimming the post, it seems to argue for reconstructing the very rigidity that LLMs let us escape, and that very aspect of LLMs is what made them useful and explode in popularity so much.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:43:37.000Z","created_at_i":1791575017,"id":50025702,"options":[],"parent_id":50025599,"points":null,"story_id":50023450,"text":"Yeah, agreed. A model or primitive on its own has no moat and frankly limited value. The paradigm behind "System One" models on the other hand is potentially huge.
https://seldon-ai.com/blog/fronter-llms-are-semantic-interpr...","title":null,"type":"comment","url":null},{"author":"porridgeraisin","children":[{"author":"outofpaper","children":[{"author":"cheesecakegood","children":[{"author":"disgruntledphd2","children":[{"author":"porridgeraisin","children":[{"author":"disgruntledphd2","children":[],"created_at":"2026-10-12T07:08:25.000Z","created_at_i":1791788905,"id":50052442,"options":[],"parent_id":50045262,"points":null,"story_id":50023450,"text":"> Their "confidence" score is rather bunk, so it's hard to say.
Then it's not calibrated (in the technical sense of the term).","title":null,"type":"comment","url":null}],"created_at":"2026-10-11T16:57:42.000Z","created_at_i":1791737862,"id":50045262,"options":[],"parent_id":50032852,"points":null,"story_id":50023450,"text":"It's calibrated where it's calibrated and not where its not. Their "confidence" score is rather bunk, so it's hard to say. As with other LLMs, you have to eval it for your setting, and if you find it calibrated, then great.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T13:37:30.000Z","created_at_i":1791639450,"id":50032852,"options":[],"parent_id":50026763,"points":null,"story_id":50023450,"text":"> If the word \u201ccalibration\u201d in probability doesn\u2019t mean anything to you, the difference isn\u2019t very apparent, but that doesn\u2019t mean it doesn\u2019t exist.
I am really sceptical that they've developed a zero shot calibrated classifier for arbitrary inputs.
That being said, if they have, then this valuation is like 100 times too low.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:27:14.000Z","created_at_i":1791581234,"id":50026763,"options":[],"parent_id":50026686,"points":null,"story_id":50023450,"text":"At least in theory (TypeSafe has been pretty close-lipped about the details so this might just be hot air, and I think the evidence is a bit spotty) this is false, since they use a different reinforcement training method.
If the word \u201ccalibration\u201d in probability doesn\u2019t mean anything to you, the difference isn\u2019t very apparent, but that doesn\u2019t mean it doesn\u2019t exist.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:19:53.000Z","created_at_i":1791580793,"id":50026686,"options":[],"parent_id":50025912,"points":null,"story_id":50023450,"text":"The are still just 1tok output of pretty standard llms just along with the logprobs converted to some json","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:01:47.000Z","created_at_i":1791576107,"id":50025912,"options":[],"parent_id":50025599,"points":null,"story_id":50023450,"text":"> Unlike full language models, I don't think the primary business of Typesafe will be serving Jev at API pricing; it'll look a lot more like putting Jev at the center of a much more expensive suite of software.
Precisely this. Should be top comment.
Also, the model moat is understated as training data for these purposes also accrues to the winner, which due to the first mover advantage as well as the distribution advantage you speak of, is typesafe. In contrast to relatively open coding data. Openai anthropic also have that, but like you say its a different business.","title":null,"type":"comment","url":null},{"author":"pantelisk","children":[{"author":"aranelsurion","children":[],"created_at":"2026-10-09T20:13:02.000Z","created_at_i":1791576782,"id":50026020,"options":[],"parent_id":50025959,"points":null,"story_id":50023450,"text":"I remember your blog post! Thanks for writing it, was pretty cool and a practical application.
Here if anyone is interested: https://pantel.is/projects/ai-gaming-companion/","title":null,"type":"comment","url":null},{"author":"overfeed","children":[],"created_at":"2026-10-09T21:02:19.000Z","created_at_i":1791579739,"id":50026505,"options":[],"parent_id":50025959,"points":null,"story_id":50023450,"text":"> [...]the fact that nobody was able to see it before typesafe is a testament of what they might be able to come up with next.
Counterpoint: there are a lot of one-hit wonders, and they vastly outnumber the idea-factory people. This is not to minimize those people, a single idea can be very successful (see Zuckerberg), but it doesn't mean your subsequent ideas will also be great (see Zuckerberg)","title":null,"type":"comment","url":null},{"author":"skeeter2020","children":[],"created_at":"2026-10-09T21:28:07.000Z","created_at_i":1791581287,"id":50026768,"options":[],"parent_id":50025959,"points":null,"story_id":50023450,"text":">> TLDR: There's a lot of value in thinking ahead and seeing the future. The clones are nice and exciting but they give me "I could have built this first, yes but you didn't" vibes. I hope they manage to keep it up and push the space forward again
This all sounds intelligent and likely, and yet we can come up with countless counter examples where the first mover is not the big winner, and nobody cares about who did it first. There is typically way more value in nailing the execution of a big idea someone else came up with, rather than "seeing the future".","title":null,"type":"comment","url":null},{"author":"user43928","children":[{"author":"senordevnyc","children":[{"author":"jeena","children":[{"author":"senordevnyc","children":[],"created_at":"2026-10-10T03:30:50.000Z","created_at_i":1791603050,"id":50029278,"options":[],"parent_id":50028214,"points":null,"story_id":50023450,"text":"I keep hearing Jev is better for accuracy, and still in my use cases Luna beats it. Latency doesn\u2019t matter to me. The cost for additional Luna tokens is negligible for my use case. Idk, I just can\u2019t figure out where to use it instead of an LLM.
But that\u2019s just for my product.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T00:22:35.000Z","created_at_i":1791591755,"id":50028214,"options":[],"parent_id":50027795,"points":null,"story_id":50023450,"text":"Did you look into self-hosted open source decision models because they are quite capable also much faster and much easier to integrate and you don't need to pay anything.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:19:09.000Z","created_at_i":1791587949,"id":50027795,"options":[],"parent_id":50026863,"points":null,"story_id":50023450,"text":"You\u2019re correct, it\u2019s a great solution for a very narrow set of high volume classification needs that require very low latency. But that\u2019s it.
I keep evaluating Jev for my product because of the hype, but the reality is that for my tasks, Luna is more accurate, only a little more expensive, and the latency doesn\u2019t matter. I\u2019d rather spend the extra $50 / month or whatever than have to shoehorn in another API and provider, and also lose the ability to change reasoning level and get reasoning summaries for eval purposes.","title":null,"type":"comment","url":null},{"author":"pantelisk","children":[{"author":"nostrebored","children":[],"created_at":"2026-10-11T00:00:58.000Z","created_at_i":1791676858,"id":50038348,"options":[],"parent_id":50029736,"points":null,"story_id":50023450,"text":"Jev does not have the properties you\u2019re talking about. It\u2019s not very good at general knowledge, it frequently approaches chance level classification on business domains, and there is so much room for someone to do better.
Doing better, I suspect, will be throwing a bigger model underneath and doing something diffusion like (we\u2019ll allocate a whole block to compute rather than just the constrained space) that doesn\u2019t compromise on latency.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T05:04:01.000Z","created_at_i":1791608641,"id":50029736,"options":[],"parent_id":50026863,"points":null,"story_id":50023450,"text":"LLMs have higher accuracy and are more versatile, but that comes with a cost (slower and more expensive to run). On the other end you can have your own mini classifier something built on top of ModernBert or DIET or even a hash classifier (these are narrow).
Models like JEV fall in the golden middle. Somewhat cheap, somewhat fast, somewhat general.
Many people have replied saying that narrow classifiers (eg bert) are better, and they are better in many ways (essentially free and faster). But... the real world is messy, in my experience it's actually very hard to make a really good classifier that will work across a specific niche domain especially when there is very fuzzy input. And the most interesting real world scenarios are nuanced and overtime there will be edge cases discovered where it fails and that will require retraining and re-evaluation etc etc. It can becomes a full project on its own that eventually collapses into a game of whackamole (improved in some direction but regressed elsewhere).\nThings like JEV (or similar models) provide ease of mind, just off load the complexity to it and move on to the next challenge type of thing.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:37:22.000Z","created_at_i":1791581842,"id":50026863,"options":[],"parent_id":50025959,"points":null,"story_id":50023450,"text":"Honestly, I don't feel the least bit of excitement here and I'm normally enthusiastic about AI.
Yes, this generates probabilities over a set of given options instead of the whole token vocabulary.
I don't see what's so exciting about it, compared to regular LLMs which support structured output options in the API.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:06:53.000Z","created_at_i":1791576413,"id":50025959,"options":[],"parent_id":50025599,"points":null,"story_id":50023450,"text":"Yes, the best way to think of a general classifier like this is like a smart switch statement. Essentially a "JEV" like thing becomes a sort of programming primitive. Once you see it, it's hard to not get excited.
But even if others surpass them and make better solutions, the fact that nobody was able to see it before typesafe is a testament of what they might be able to come up with next.
I sound like a fanboy but I swear I 'm unaffiliated with typesafe. I was building my own version of this way before they announced JEV (mine was ALE and it was mentioned here on HN for a bit), in use for VR gaming (so one can give commands to NPCs with voice and supports multiple commands in sequence in a single pass), but I missed the "killer usecase" of being a new primitive, like everyone else.
TLDR: There's a lot of value in thinking ahead and seeing the future. The clones are nice and exciting but they give me "I could have built this first, yes but you didn't" vibes. I hope they manage to keep it up and push the space forward again","title":null,"type":"comment","url":null},{"author":"nico","children":[{"author":"vrighter","children":[],"created_at":"2026-10-10T06:43:02.000Z","created_at_i":1791614582,"id":50030333,"options":[],"parent_id":50027000,"points":null,"story_id":50023450,"text":"i read that quote as "how it looks is more important than if it works"","title":null,"type":"comment","url":null},{"author":"Darmani","children":[{"author":"dannyw","children":[{"author":"kolinko","children":[{"author":"behnamoh","children":[],"created_at":"2026-10-11T02:36:26.000Z","created_at_i":1791686186,"id":50039284,"options":[],"parent_id":50031335,"points":null,"story_id":50023450,"text":"To be fair, there are still many jailbreaks for instruction tuned models. Surely less than before, but it could be because models got smarter (Claude often "reasons" that something is fishy even if it appears in the prompt or appears like a system prompt).","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T10:05:29.000Z","created_at_i":1791626729,"id":50031335,"options":[],"parent_id":50030917,"points":null,"story_id":50023450,"text":"I think the main benefit of instruction finetuning was that the model is less likely to get prompt injections and more likely to follow the specified format?
Pure davinci was easily confused by what was system prompt, user prompt and its own answers.","title":null,"type":"comment","url":null},{"author":"behnamoh","children":[],"created_at":"2026-10-11T02:34:27.000Z","created_at_i":1791686067,"id":50039269,"options":[],"parent_id":50030917,"points":null,"story_id":50023450,"text":"I remember those days and I'm glad to see another one who's been in the industry since the very beginning :) I was quite sad when davinci-003 was taken away; still remember the green-yellow-orange-red highlights the OpenAI console would make for log probabilities of tokens and all the cool things we could do with a pretrained model.
In my courses, I teach students about that era by using gemma 4 (pretrained) and showing how you can emulate a conversation even on something that wasn't post-trained for chat.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T08:35:53.000Z","created_at_i":1791621353,"id":50030917,"options":[],"parent_id":50030757,"points":null,"story_id":50023450,"text":"It was impressive, but not that much. If you had a basic prompt prefix like:
\u201cThe following is a conversation between a human and a helpful AI assistant.
Human: [prompt]\\n\nAI:\u201d
You would get consistent conversational chat, with the main limit being the model\u2019s limited context window, and obviously frontier intelligence at the time.
I played around quite a bit with davinci-003 and conversational systems before ChatGPT. I kept using this format and API for a while, but at least during the \u2018free research preview\u2019 era, found it generally smarter than chatgpt. 3.5-turbo was probably the watershed moment where I moved away from completions on base models; to chat completions.
If you\u2019d like to emulate this experience, spin up a base (not instruction tuned) checkpoint of Llama 1; or a more recent base model. You might underestimate how much \u2018intelligence\u2019 you can get :)
Back then the API exposed a lot of controls from sampling to logits, so you could specify a custom stop token (some rarely used Unicode); then switch to near-greedy sampling for more reliable \u201cfunction calls\u201d, etc; and then switch back to sampling for chat.
In the early days, before releasing something with the API, you had to get your use case approved by OpenAI, like the App Store review process.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T08:03:41.000Z","created_at_i":1791619421,"id":50030757,"options":[],"parent_id":50027000,"points":null,"story_id":50023450,"text":"The instruction tuning between 3 and 3.5 was a big deal. If you prompted raw GPT 3 "Tell me a story about a mouse", it was as likely to continue "2) Tell me a story about a cow. 3) Tell me a story about a rabbit" as it was to tell you a story. I tried to use GPT 3 for work and found its only utility was to get over blank page fear by writing something so bad I would fix it in anger.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:48:56.000Z","created_at_i":1791582536,"id":50027000,"options":[],"parent_id":50025599,"points":null,"story_id":50023450,"text":"> But the model itself matters less than the product experiences you build around the model
This is a very important insight. And it applies to LLMs as well. Very few people were impressed with the capabilities of GPT 3, it was mostly a techie novelty
But then when they added chat on top of gpt 3.5, all of a sudden it was a huge hit. Sure there were improvements in the model from 3 to 3.5, but the biggest impact was from the chat experience
Conversely, when they created Eliza, a basic chatbot more than 50 years ago, people even got addicted to it, despite it\u2019s ai model being something super rudimentary and basic compared to what we have now. The model capabilities didn\u2019t matter as much as the experience the chat created","title":null,"type":"comment","url":null},{"author":"olalonde","children":[{"author":"porridgeraisin","children":[{"author":"zhivota","children":[{"author":"majormajor","children":[{"author":"sks_15","children":[],"created_at":"2026-10-10T04:32:37.000Z","created_at_i":1791606757,"id":50029580,"options":[],"parent_id":50029032,"points":null,"story_id":50023450,"text":"It really shines when you use that decision routing in industries where they need to automate these routings reliably with very low latency, higer certainly and ,more important, at scale. Take call centers for example. If a voice company needs to automate escalation etc. an llm would be an overkill. That's where system 1 models shine.","title":null,"type":"comment","url":null},{"author":"porridgeraisin","children":[{"author":"disgruntledphd2","children":[{"author":"porridgeraisin","children":[{"author":"disgruntledphd2","children":[],"created_at":"2026-10-10T17:18:36.000Z","created_at_i":1791652716,"id":50035019,"options":[],"parent_id":50032888,"points":null,"story_id":50023450,"text":"Yeah, it's not the development but the maintenance that kills you in both software and other information fields.
The whole text to SQL thing is about data teams being a bottleneck to the business (rather like software people). The goal is good in theory but honestly an Excel plugin is probably a better solution in many cases.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T13:41:53.000Z","created_at_i":1791639713,"id":50032888,"options":[],"parent_id":50032819,"points":null,"story_id":50023450,"text":"> I saw a place which essentially tasked the data analysts/scientists in an area to describe the good tables, and assumptions and necessary context to make decent queries/analyses, and that definitely lead to much, much better (almost usable) results.
Yep. After this has been done, the later text to sql outputs works half decent. The major issue I heard was that whenever a small assumption changes few months down the line, while humans are able to "JIT" apply it to each of their queries, these models sometimes lose it in all of the context and so it becomes hard to progressively update anything - you end up having to redo the whole workflow.
And in my experience, even the more complex dashboards requested by product or upper mgmt were stood up within the day by humans. Or the next day (but that was cos some ETL would be scheduled overnight rather than immediately). And I am not sure if speeding up this part would even matter to the guy who requested the dashboard. So I am not sure if this is a major bottleneck people should try to innovate in. If openai does the work for you then great but otherwise. Humans aided by copilot-of-old style autocomplete is probably more than enough.
Needless to say, text to sql is pointless for application queries made during API calls.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T13:35:08.000Z","created_at_i":1791639308,"id":50032819,"options":[],"parent_id":50030580,"points":null,"story_id":50023450,"text":"> These guys also love text to sql. I know 4 people each implementing text to sql for their separate companies.
Text to SQL is a really bad idea, as the agent will just generate really convincing and terrible analyses that lack all business context.
I mean, you could implement some kind of semantic layer, but that's a lot of work, for uncertain reward.
I saw a place which essentially tasked the data analysts/scientists in an area to describe the good tables, and assumptions and necessary context to make decent queries/analyses, and that definitely lead to much, much better (almost usable) results.
As I have said many times, what the world really needs is SQL2Text, rather than Text2SQL.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T07:29:35.000Z","created_at_i":1791617375,"id":50030580,"options":[],"parent_id":50029032,"points":null,"story_id":50023450,"text":"It is. But the thing is, that is also a lot of the usage for "AI" these days. LLM based search pipelines are being deployed but not really used. Chatbots are a great hit. And then of course, in coding, finance, sales, accounts and audit, etc, it is being used to accelerate output.
These guys also love text to sql. I know 4 people each implementing text to sql for their separate companies. Or text-to-redash dashboard in some cases. Funnily enough, text to sql is one of the things where LLMs are clearly sub-human. Mostly due to lack of easy verifiability.","title":null,"type":"comment","url":null},{"author":"drob518","children":[],"created_at":"2026-10-10T12:56:32.000Z","created_at_i":1791636992,"id":50032469,"options":[],"parent_id":50029032,"points":null,"story_id":50023450,"text":"If it was simple they would just use an LLM to write code to do it. These decisions are \u201cfuzzy\u201d in the sense that they are difficult to make in just code. They need the \u201csmarts\u201d of a language model to evaluate something, but the final answer is basically trivial (e.g. choose one of these five options, not write me 20 page of text or one shot this whole complex program).
There are a lot of those types of problems in the world, replacing what was previously spotty hardcoded decision logic that worked 70% of the time with something that now gets it right 95% of the time. Even if not perfect, the fact that you can dramatically close the gap means a lot lower cost overall, so it\u2019s worth it.","title":null,"type":"comment","url":null},{"author":"zhivota","children":[],"created_at":"2026-10-11T10:21:09.000Z","created_at_i":1791714069,"id":50041585,"options":[],"parent_id":50029032,"points":null,"story_id":50023450,"text":"It's pretty high value add when you consider it's replacing human beings assessing each case. This category of thing is present all over enterprises. It was outsourced to BPO shops in India and the Philippines before, now it's going to AI.
But you're right that more complex solutions are still a bit elusive. I use a bot to plan my strength workouts based on my soreness level and what equipment I have available and it's decent, but not perfect.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T02:47:19.000Z","created_at_i":1791600439,"id":50029032,"options":[],"parent_id":50028707,"points":null,"story_id":50023450,"text":"A lot of that seems like low-value-add/low-margin "simple" automation.
I haven't personally yet found a net-new-capability bigger than that from LLMs here.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T01:48:53.000Z","created_at_i":1791596933,"id":50028707,"options":[],"parent_id":50027676,"points":null,"story_id":50023450,"text":"Right, example being expense pre-approvals (for low dollar value purchases like office equipment). General purpose LLM not required, and the product experience surrounding it does matter (reporting, auditing, fine tuning future responses, etc.).","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:04:19.000Z","created_at_i":1791587059,"id":50027676,"options":[],"parent_id":50027540,"points":null,"story_id":50023450,"text":"If youre the one or have spoken to someone implementing "AI solutions" inside large companies recently, a decent chunk of it is soft policy enforcement with very basic context. They moved from gemini 2.5 flash lite type models to jev. Which is also why I found the price comparisons to "GPT Astra" on Twitter rather funny.","title":null,"type":"comment","url":null},{"author":"reexpressionist","children":[{"author":"SwellJoe","children":[{"author":"dvfjsdhgfv","children":[],"created_at":"2026-10-10T18:10:23.000Z","created_at_i":1791655823,"id":50035544,"options":[],"parent_id":50030900,"points":null,"story_id":50023450,"text":"The parent wasn't giving an answer but adding a more nuance which I find valuable.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T08:33:44.000Z","created_at_i":1791621224,"id":50030900,"options":[],"parent_id":50028266,"points":null,"story_id":50023450,"text":"That's a lot of words to not answer the question.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T00:34:39.000Z","created_at_i":1791592479,"id":50028266,"options":[],"parent_id":50027540,"points":null,"story_id":50023450,"text":"A major limitation of directly using "System 1 decision models" (a.k.a., logistic regression, and related uncalibrated classifiers over the output/logit space) in enterprise settings (or other high-stakes settings) is that such estimators are not reliable estimators of the predictive uncertainty in the presence of covariate shifts, and such estimators also lack a means of instance-wise data attribution (i.e., interpretability-by-exemplar), so they're not the ideal estimator for verification, routing, uncertainty over retrieval and tool-calls, etc.
For decision-making with neural networks, we instead need the older idea of estimators of the predictive uncertainty with constraints in the feature-representation space (over training/support of the estimator), as with Similarity-Distance-Magnitude estimators: https://pypi.org/project/reexpress-sdm/","title":null,"type":"comment","url":null},{"author":"rebyn","children":[{"author":"SalariedSlave","children":[{"author":"locknitpicker","children":[],"created_at":"2026-10-10T16:36:07.000Z","created_at_i":1791650167,"id":50034541,"options":[],"parent_id":50030511,"points":null,"story_id":50023450,"text":"> the claim is extraordinary, yet the answers cover only the mundane.
I wonder how many of these posts are astroturfed.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T07:17:55.000Z","created_at_i":1791616675,"id":50030511,"options":[],"parent_id":50030206,"points":null,"story_id":50023450,"text":"it's not just you.
the claim is extraordinary, yet the answers cover only the mundane.","title":null,"type":"comment","url":null},{"author":"iinnPP","children":[{"author":"72deluxe","children":[],"created_at":"2026-10-10T12:43:29.000Z","created_at_i":1791636209,"id":50032381,"options":[],"parent_id":50030815,"points":null,"story_id":50023450,"text":"I can attest that the quality of pizzas has risen.","title":null,"type":"comment","url":null},{"author":"its-summertime","children":[],"created_at":"2026-10-10T13:01:25.000Z","created_at_i":1791637285,"id":50032522,"options":[],"parent_id":50030815,"points":null,"story_id":50023450,"text":"Isn't that just normal for Hacker News?","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T08:15:41.000Z","created_at_i":1791620141,"id":50030815,"options":[],"parent_id":50030206,"points":null,"story_id":50023450,"text":"I just wanted to point out that I have started to note this "phenomenon", seemingly globally, on any topic.
The sum of people not even close to the topic has risen, according to my feels at least.
Has this rang true for anyone else?","title":null,"type":"comment","url":null},{"author":"PunchyHamster","children":[],"created_at":"2026-10-10T09:06:43.000Z","created_at_i":1791623203,"id":50031042,"options":[],"parent_id":50030206,"points":null,"story_id":50023450,"text":"Because answering the why reduces investor hype train","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T06:23:23.000Z","created_at_i":1791613403,"id":50030206,"options":[],"parent_id":50027540,"points":null,"story_id":50023450,"text":"Is it me or so far all the current HN replies to this particular \u201cwhy\u201d ask seem not even remotely answering it?","title":null,"type":"comment","url":null},{"author":"adilkhanovkz","children":[],"created_at":"2026-10-10T12:58:25.000Z","created_at_i":1791637105,"id":50032491,"options":[],"parent_id":50027540,"points":null,"story_id":50023450,"text":"I think because pretty much every step a business takes is an allocation decision: where to put money, people, time. But it will really pay off only when the models of companies themselves become probabilistic, and the decisions are explainable, at least to some degree.","title":null,"type":"comment","url":null},{"author":"ghm2180","children":[{"author":"nowittyusername","children":[{"author":"nostrebored","children":[],"created_at":"2026-10-10T23:54:58.000Z","created_at_i":1791676498,"id":50038325,"options":[],"parent_id":50033653,"points":null,"story_id":50023450,"text":"Yeah agreed that vision is a nice fit here. I\u2019ve gotten some really impressive results by just throwing xgboost on 25 classification related questions.
System one models do not have an obvious use in voice ai, unless you are choosing from a list of potential utterances (eg embed -> top k -> choose from top k)","title":null,"type":"comment","url":null},{"author":"ghm2180","children":[{"author":"nowittyusername","children":[],"created_at":"2026-10-11T16:41:32.000Z","created_at_i":1791736892,"id":50045090,"options":[],"parent_id":50043018,"points":null,"story_id":50023450,"text":"Yes, i looked in to pipecat stuff a while ago and revisit it often because its relevant for voice agents, but i never ended up using any of their stuff as my custom solutions were better. Also id like to hint that smart end of turn detection is not needed at all in a cascaded system IF you have access to the inference engine of the llm you are using in the sandwich. I dont want to reveal it all but lets just say its no longer a headache or any issue at all. If using 3d party api then yes it helps to have a smart eot system. But i try to keep everything fully local.","title":null,"type":"comment","url":null}],"created_at":"2026-10-11T13:15:07.000Z","created_at_i":1791724507,"id":50043018,"options":[],"parent_id":50033653,"points":null,"story_id":50023450,"text":"Thanks. I am curious were you trying something like pipecat's Turn finalization? https://docs.pipecat.ai/api-reference/server/utilities/turn-...","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T15:02:10.000Z","created_at_i":1791644530,"id":50033653,"options":[],"parent_id":50032938,"points":null,"story_id":50023450,"text":"That was my first insight as well. i have been working on a voice agent for a while and decided to spend about 4 days messing with system one models to see where it could fit in my cascaded system. after 4 days i ripped it out of my system as i found better solutions.. i tried really hard to find a use for it but every time i found there was no need for it so i stopped trying to shoehorn it in. i also spent time with the vision side of things for this thing, and i think that where the biggest gains will be IMO. making decision based on visual ques and in pixel space only was the biggest win.","title":null,"type":"comment","url":null},{"author":"yed","children":[],"created_at":"2026-10-10T18:49:49.000Z","created_at_i":1791658189,"id":50035902,"options":[],"parent_id":50032938,"points":null,"story_id":50023450,"text":"JEV only sees transcribed text, so long term I don\u2019t see this use case beating audio models built for this task, or especially voice to voice models.","title":null,"type":"comment","url":null},{"author":"nostrebored","children":[{"author":"ghm2180","children":[],"created_at":"2026-10-11T13:10:41.000Z","created_at_i":1791724241,"id":50042984,"options":[],"parent_id":50038312,"points":null,"story_id":50023450,"text":"Pipecat already does this. It does it for end of turn detection and it's pretty damn good. It's just that it calls the same LLM I have to do it, potentially expensive and prone to latency and not calibrated.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T23:52:33.000Z","created_at_i":1791676353,"id":50038312,"options":[],"parent_id":50032938,"points":null,"story_id":50023450,"text":"This is really not a good idea.
Modern stt is baking semantic EOT into the prediction. It makes sense. Conversation has many potential ends, and you have to deal with misprediction of current words that may signal completion, but mel spectrograms contain a ton of information as well.
In your version, we\u2019d have to deal with partial transcriptions and a system that is not natively audio aware. It would also be much more expensive.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T13:45:55.000Z","created_at_i":1791639955,"id":50032938,"options":[],"parent_id":50027540,"points":null,"story_id":50023450,"text":"The simplest example could be replacement or at the least an economical complement to the LLM step in cascading VoiceAI, almost every start of the turn a local JEV could make a decision for end of user turn detection:\n- is the user finished with their thought? If so delegate to next system(LLM another JEV)\n- Should I wait for the user to complete their thought for a few seconds? Set a timer and when time out expires, ask the user to continue.
Like just tuning this model can make it cheaper for voice AI providers to do semantic turn detection over the space of responses.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:47:25.000Z","created_at_i":1791586045,"id":50027540,"options":[],"parent_id":50025599,"points":null,"story_id":50023450,"text":"> Decision models have the potential to have an even larger impact on the Real World than LLMs have to this point
Why?","title":null,"type":"comment","url":null},{"author":"baxtr","children":[],"created_at":"2026-10-10T04:43:48.000Z","created_at_i":1791607428,"id":50029644,"options":[],"parent_id":50025599,"points":null,"story_id":50023450,"text":"I think you aptly describe the benefits this approach can offer.
What I haven\u2019t seen answered is why it isn\u2019t reproducible within very short amount of time and low effort.
Sure, the large labs are the lazy incumbents at this stage. But any other startups would be able to copy the experience. What\u2019s the barrier of entry that I am missing?","title":null,"type":"comment","url":null},{"author":"zqy123007","children":[],"created_at":"2026-10-10T09:59:23.000Z","created_at_i":1791626363,"id":50031309,"options":[],"parent_id":50025599,"points":null,"story_id":50023450,"text":"> There's the potential for an inverse LLM play.
Exactly.
If JEV is here to stay, we can have another split along the "one LLM rule them all" regime
name it worker or implementator or something
- mostly for agent with with spec (not directly for human, as we tend be handwave with vague intent)\n- follows the instruction verbatim\n- lives in sandbox by default\n- basically sol6; but WITHOUT magical situational awareness, lets work in group BS","title":null,"type":"comment","url":null},{"author":"acjohnson55","children":[{"author":"kijin","children":[{"author":"echelon","children":[],"created_at":"2026-10-10T15:11:49.000Z","created_at_i":1791645109,"id":50033734,"options":[],"parent_id":50032273,"points":null,"story_id":50023450,"text":"Still, what a bold early stage bet.
I suppose the team and the surprise warrant that.","title":null,"type":"comment","url":null},{"author":"acjohnson55","children":[],"created_at":"2026-10-11T01:48:29.000Z","created_at_i":1791683309,"id":50039029,"options":[],"parent_id":50032273,"points":null,"story_id":50023450,"text":"For sure. It still seems pretty speculative, but that's the game.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T12:29:47.000Z","created_at_i":1791635387,"id":50032273,"options":[],"parent_id":50032108,"points":null,"story_id":50023450,"text":"> but I think it's a big leap to say that they will be able to dominate the category.
Which is why they are valued at "only" 7.5B, while frontier labs with general-purpose LLMs have valuations in the order of trillions.
Horowitz et al. invest in probabilities, not certainty. If Typesafe doesn't make it, they'll have other horses in the race too.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T12:09:23.000Z","created_at_i":1791634163,"id":50032108,"options":[],"parent_id":50025599,"points":null,"story_id":50023450,"text":"> Decision models have the potential to have an even larger impact on the Real World than LLMs have to this point (which is obviously quite large). But the model itself matters less than the product experiences you build around the model, and its very likely that the incumbent labs are treating the area as something more like "oh yeah I guess we can ship that and then forget about it" rather than investing in what building business processes on decision models looks like.
I agree that the potential is enormous, but I'm not sure why it would be Typesafe who figures out the best way to develop and package this capability. They have a bit of an advantage in the amount of time they have been thinking about this problem specifically, but I think it's a big leap to say that they will be able to dominate the category.","title":null,"type":"comment","url":null},{"author":"sunaookami","children":[],"created_at":"2026-10-10T16:59:25.000Z","created_at_i":1791651565,"id":50034804,"options":[],"parent_id":50025599,"points":null,"story_id":50023450,"text":">Decision models have the potential to have an even larger impact on the Real World than LLMs have to this point
They really haven't.","title":null,"type":"comment","url":null},{"author":"make3","children":[],"created_at":"2026-10-10T20:40:38.000Z","created_at_i":1791664838,"id":50036949,"options":[],"parent_id":50025599,"points":null,"story_id":50023450,"text":"> it'll look a lot more like putting Jev at the center of a much more expensive suite of software
but anyone can do that, that's not their product, Jev is","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:34:37.000Z","created_at_i":1791574477,"id":50025599,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Decision models have the potential to have an even larger impact on the Real World than LLMs have to this point (which is obviously quite large). But the model itself matters less than the product experiences you build around the model, and its very likely that the incumbent labs are treating the area as something more like "oh yeah I guess we can ship that and then forget about it" rather than investing in what building business processes on decision models looks like. Unlike full language models, I don't think the primary business of Typesafe will be serving Jev at API pricing; it'll look a lot more like putting Jev at the center of a much more expensive suite of software.
There's the potential for an inverse LLM play. In contrast with LLMs, all that seems to matter is the model, and the products the labs build around the models are all really samey and boring; the same left panel list of agents, main view agent conversation, right hand extra context, and we're now in the era of everyone creating the same cutesey furry friend on top of all this tech.","title":null,"type":"comment","url":null},{"author":"moralestapia","children":[],"created_at":"2026-10-09T19:35:59.000Z","created_at_i":1791574559,"id":50025618,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Nothing beats nepo, brother.","title":null,"type":"comment","url":null},{"author":"hkalbasi","children":[],"created_at":"2026-10-09T19:39:34.000Z","created_at_i":1791574774,"id":50025656,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"> OpenAI's own Decisions API [1] beats it
Jev is 42$/B but OpenAI is 100$/B token.","title":null,"type":"comment","url":null},{"author":"bushbaba","children":[],"created_at":"2026-10-09T19:53:57.000Z","created_at_i":1791575637,"id":50025811,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"A major VC could type safe ai money, then head to a larger AI company looking to raise their series E+ and demand they acquire typesafe as part of their funding allotment.
such an arrangement can end up beneficial to the VC firm","title":null,"type":"comment","url":null},{"author":"throwaw12","children":[],"created_at":"2026-10-09T20:11:35.000Z","created_at_i":1791576695,"id":50026008,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"You are right in terms of how fast competition created alternatives.
But, for OpenAI this is not a primary business, for open source models as well, so they will not be chasing the market and customers to buy their product and promise them to maintain it.
TypeSafe will do all this, they will try to understand your use cases and then solve your pain point, while others are providing raw material.","title":null,"type":"comment","url":null},{"author":"jgilias","children":[{"author":"phalangion","children":[],"created_at":"2026-10-09T20:25:58.000Z","created_at_i":1791577558,"id":50026153,"options":[],"parent_id":50026128,"points":null,"story_id":50023450,"text":"What\u2019s the difference?","title":null,"type":"comment","url":null},{"author":"hbrn","children":[{"author":"jgilias","children":[{"author":"hbrn","children":[{"author":"shados","children":[{"author":"hbrn","children":[{"author":"jgilias","children":[{"author":"hbrn","children":[],"created_at":"2026-10-10T16:22:35.000Z","created_at_i":1791649355,"id":50034401,"options":[],"parent_id":50030236,"points":null,"story_id":50023450,"text":"> does this advice correspond to our policy docs
This is not a \u201csystem one\u201d question.
You\u2019d be much better off using a proper LLM here: more accurate, can provide justification, can be steered when mistakes are made. Jev is going to be a coin flipper with almost no knobs.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T06:28:06.000Z","created_at_i":1791613686,"id":50030236,"options":[],"parent_id":50028001,"points":null,"story_id":50023450,"text":"That\u2019s circular, sure. But if we think about potential real-world tasks where someone might, say, use it to classify on \u201cdoes this advice correspond to our policy docs\u201d, you\u2019d absolutely push the answer (the policy docs) into its context..
What am I missing?
As in, neither LLMs nor Jev are truth engines. Truth comes from the provided context.. Plus weights.. kind of fuzzy, sorry I\u2019m thinking out \u201cloud\u201d","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:50:22.000Z","created_at_i":1791589822,"id":50028001,"options":[],"parent_id":50027520,"points":null,"story_id":50023450,"text":"> give it a list of probability for each number and ask it whats the probability of each number
Did i hear that correctly? In order for Jev to be accurate you have to give it the answer before asking for the answer?
(btw this is exactly how Jev is playing games).","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:45:28.000Z","created_at_i":1791585928,"id":50027520,"options":[],"parent_id":50027235,"points":null,"story_id":50023450,"text":"The dice roll prediction is about the way the prompt is setup misunderstanding how Jev works (they treat the confidence score as a probability score, which it isn't).
If you instead give it a list of probability for each number and ask it whats the probability of each number, the result will be accurate.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:12:57.000Z","created_at_i":1791583977,"id":50027235,"options":[],"parent_id":50026808,"points":null,"story_id":50023450,"text":"Yeah, but I wouldn't be surprised OpenAI's decision API is a post-trained Luna with confidence calibration.
Typesafe claims that Jev is calibrated, but there are plenty of examples where it completely fails (predicting die roll being the most obvious one).
Unfortunately calibration is hard to benchmark.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:33:02.000Z","created_at_i":1791581582,"id":50026808,"options":[],"parent_id":50026521,"points":null,"story_id":50023450,"text":"Thanks, fixed my understanding!
Do you think though that Luna being a model post-trained for chat produces over-confidence in logprobs?","title":null,"type":"comment","url":null},{"author":"adrian17","children":[{"author":"hbrn","children":[],"created_at":"2026-10-09T22:34:15.000Z","created_at_i":1791585255,"id":50027426,"options":[],"parent_id":50027282,"points":null,"story_id":50023450,"text":"Jev API returns both confidence and probabilities.
But confidence value is just a function applied to probabilities. It is not coming from the model, and it carries no additional information.
It is documented btw, and yet you will see plenty of claims that Jev is better than LLM because it returns both.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:17:33.000Z","created_at_i":1791584253,"id":50027282,"options":[],"parent_id":50026521,"points":null,"story_id":50023450,"text":"Maybe I'm missing something, but why couldn't it be generated by the model? In older classification tasks with transformers like BERT, you could absolutely obtain a confidence score.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:03:54.000Z","created_at_i":1791579834,"id":50026521,"options":[],"parent_id":50026128,"points":null,"story_id":50023450,"text":"And what do you think Jev confidence score is?
Here's a hint: confidence is not generated by a model.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:22:55.000Z","created_at_i":1791577375,"id":50026128,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Isn\u2019t the OpenAI decisions API basically just Luna cosplaying a decisions model and pretending the confidence score isn\u2019t just a hallucination?","title":null,"type":"comment","url":null},{"author":"girvo","children":[],"created_at":"2026-10-09T20:41:27.000Z","created_at_i":1791578487,"id":50026329,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Counterpoint: my work has already allowed us to call and test Jev. Those others? Who knows when, if ever.","title":null,"type":"comment","url":null},{"author":"user3939382","children":[],"created_at":"2026-10-09T20:46:31.000Z","created_at_i":1791578791,"id":50026376,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Investments aren\u2019t made because the product is amazing, they\u2019re made because there\u2019s a compelling exit scenario. Engineers don\u2019t want to hear this but more generally, the critical success factors for a business aren\u2019t product or engineering they\u2019re relationships i.e. sales and team dynamics. If technical excellence dictated business outcomes in tech Salesforce wouldn\u2019t exist for example.","title":null,"type":"comment","url":null},{"author":"nico","children":[],"created_at":"2026-10-09T21:44:19.000Z","created_at_i":1791582259,"id":50026956,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Yup, I also released an open source classifiers tool, Jeffy. It comes with 68 pre trained classifiers which run and train on CPU alone. They run locally and are faster than Jev/Laya/Decisions. And they can do things like label email, all the way to even playing Doom
* https://playground.jeffyclassify.com/#doom
* https://github.com/nicobrenner/jeffy","title":null,"type":"comment","url":null},{"author":"rpdillon","children":[],"created_at":"2026-10-09T21:50:56.000Z","created_at_i":1791582656,"id":50027019,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"In my experience with OpenAI's decisions endpoint, it tends to return either 0 or 1 and doesn't return middle confidence levels very much at all. Would be interested to hear if others have experienced the same.","title":null,"type":"comment","url":null},{"author":"shados","children":[{"author":"nacs","children":[],"created_at":"2026-10-09T22:47:51.000Z","created_at_i":1791586071,"id":50027543,"options":[],"parent_id":50027489,"points":null,"story_id":50023450,"text":"Hm? Jev's actual (network) latency is not that great and even small local models are doing far better latency-wise.
The Microsoft article GP linked even shows the MS model having 95ms latency.
Even on price Jev being matched (the same MS model is "Input tokens cost $0.042 USD per million tokens. Output tokens are free.", same as Jev).
Cloudflare's Clef-flash model is actually slightly cheaper: "$0.038 in / $0 out per 1M" too.
Jev is being matched or exceeded in performance and price within a month of them going public.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:40:44.000Z","created_at_i":1791585644,"id":50027489,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"I honestly was worried for them. Now, even with all the clones, they generally still come up on top in price and latency, but "good enough" is often sufficient.
Guess the (investment) market has spoken.","title":null,"type":"comment","url":null},{"author":"shmoogy","children":[{"author":"lifeisloving","children":[{"author":"jofzar","children":[],"created_at":"2026-10-10T03:30:45.000Z","created_at_i":1791603045,"id":50029277,"options":[],"parent_id":50027737,"points":null,"story_id":50023450,"text":"That's why it's useful, it's for when you don't know what needs to be classified or if you want to offer this classification models to your users.","title":null,"type":"comment","url":null},{"author":"nlpnerd","children":[],"created_at":"2026-10-10T05:04:23.000Z","created_at_i":1791608663,"id":50029738,"options":[],"parent_id":50027737,"points":null,"story_id":50023450,"text":"Because training your own model requires substantial effort around data annotations, and then dealing with data drift etc in production.","title":null,"type":"comment","url":null},{"author":"baq","children":[],"created_at":"2026-10-10T10:46:37.000Z","created_at_i":1791629197,"id":50031576,"options":[],"parent_id":50027737,"points":null,"story_id":50023450,"text":"It\u2019s not training that\u2019s the hard part, it\u2019s when you don\u2019t know your use case or it keeps changing all the time","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:11:40.000Z","created_at_i":1791587500,"id":50027737,"options":[],"parent_id":50027570,"points":null,"story_id":50023450,"text":"I just dont understand why anyone needs a general purpose classifier. Just build a classifier for your specific use cases.
Fortunately Jev is cheap, so I dont think it matters too much, but I think its robbing people of the opportunity to learn and implement this themselves.
Also, I dont really want 3 companies responsible for censorship/classification.","title":null,"type":"comment","url":null},{"author":"nwienert","children":[{"author":"andrewingram","children":[],"created_at":"2026-10-10T11:03:39.000Z","created_at_i":1791630219,"id":50031694,"options":[],"parent_id":50030411,"points":null,"story_id":50023450,"text":"I had the opposite experience. Clef was worse in accuracy, and an order of magnitude slower.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T06:57:39.000Z","created_at_i":1791615459,"id":50030411,"options":[],"parent_id":50027570,"points":null,"story_id":50023450,"text":"Hm I had built out a system that tested a few hundred decisions and clef beat Jev pretty handily, I think there was actually 0 instances where clef regressed and it improved about 20% of the false decisions.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:51:58.000Z","created_at_i":1791586318,"id":50027570,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Jev outperforms clef and OpenAI decisions on most of my tasks, outside of when I need multi modal image going into it. Jev is also cheaper but costs are minimal overall.","title":null,"type":"comment","url":null},{"author":"charm137","children":[{"author":"notfromhere","children":[],"created_at":"2026-10-09T23:30:03.000Z","created_at_i":1791588603,"id":50027871,"options":[],"parent_id":50027807,"points":null,"story_id":50023450,"text":"An Enterprise client at that size can just be someone at an f500 put a credit card in","title":null,"type":"comment","url":null},{"author":"9dev","children":[],"created_at":"2026-10-10T07:46:48.000Z","created_at_i":1791618408,"id":50030674,"options":[],"parent_id":50027807,"points":null,"story_id":50023450,"text":"Currently you can\u2019t even add a VAT ID to your account, leading to invalid invoices for all EU customers.
So yeah, not too much enterprise sales experience there for sure.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:20:53.000Z","created_at_i":1791588053,"id":50027807,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"I see Typeface AI have 30-40 people working for them on LinkedIn (some are VC advisors / board members), a lot of them being engineers. Their openings suggest a strong developer market focus (to begin with). I don't necessarily see a lot of people at the company with enterprise sales channel experience, but they already claim to have several Fortune 500 companies as clients (maybe the VCs are helping there or there is natural dev-driven traction).
So they clearly have a product, a strategy around it and perhaps the compliance scaffolding (SOC2 Type II etc) that may be needed before actually being able to charge money for it. They also have the right brand names associated with the founding team. Execution, so far, seems good enough to create a splash, at least.
As an investor, the question(s) to ask is (in my view): "How do they make money? Will that way to make money survive?". The answer to the first: selling input tokens and perhaps subscriptions/credits eventually.The answer to the second: "Yes, but with the risk of unit revenues declining faster than their unit costs". How can they mitigate this problem: by being big (scale / mindshare etc) so that their unit costs (including for customer acq) fall faster than their unit revenues will - I believe that is the question most AI companies are trying to tackle these days. Any new competitor will have to tackle basic fixed costs (of getting started) first before even getting to the stage of having the luxury of worrying about unit-economics.
So yes, they might eventually be competed away but whoever is in their team is trying hard to make a useful product/ecosystem and that should be applauded, not ridiculed with "it's all marketing". This is way more than a simple github/huggingface-based open-source replica solution can hope to achieve without institutional backing (either big-tech or system-integrators).
What should rightly be questioned, of course, are the valuations the VCs are providing to them in hopes of passing this hot potato to a willing buyer (say a hardware maker like NVidia) - the incentives there are very well defined and depend very much on perceived TAM (which lately is on very shaky ground given how far token pricing has fallen causing, among other things, OpenAI to "miss" on the market's expectations for annualized revenues, even before they're listed!) [1]
[1]: https://www.ft.com/content/b66a9858-f8fb-46cb-b506-44bfe26fc...","title":null,"type":"comment","url":null},{"author":"neilellis","children":[{"author":"phoghed","children":[],"created_at":"2026-10-10T02:18:49.000Z","created_at_i":1791598729,"id":50028887,"options":[],"parent_id":50028525,"points":null,"story_id":50023450,"text":"Yeah, tried a couple open source ones, same experience. Seems to me a ton of people are in a mad dash to create one of these and most seem to have over fit on the Jev benchmark to get some claim to fame of beating them or replicating their couple years of work in two days or whatever.
Then everyone breathlessly repeats this story about how Jev is useless because open source models they never tried claim to do the same thing and better.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T01:17:09.000Z","created_at_i":1791595029,"id":50028525,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Did you actually try them, I tried a sample they were rubbish compared to Jev - I'm not saying they will survive, but the competition CURRENTLY sucks - and also they aren't actually cheaper if you need them at scale. OpenAI is coming in twice the cost and so is Cloudflare. Ad yes you can finetune, and yes I do, but it also SUCKS. Fine-tuning and managing your own datasets is yet another timesink. So we'll see, but it's not quite black and white.","title":null,"type":"comment","url":null},{"author":"fastball","children":[],"created_at":"2026-10-10T03:00:11.000Z","created_at_i":1791601211,"id":50029104,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"And there are dozens of open-weight LLMs, but OpenAI and Anthropic are both immensely valuable because their models are actually intelligent.","title":null,"type":"comment","url":null},{"author":"jofzar","children":[],"created_at":"2026-10-10T03:15:40.000Z","created_at_i":1791602140,"id":50029194,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Testing I have seen from internal teams have shown it's still the best and cheapest model so far for pure classification work.","title":null,"type":"comment","url":null},{"author":"vrighter","children":[],"created_at":"2026-10-10T06:41:54.000Z","created_at_i":1791614514,"id":50030325,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"because all that is needed is to delete the outer while loop. it's an llm that is only run once and generates just one token. But it outputs confidence levels! Yeah confidence levels are the only output from an llm","title":null,"type":"comment","url":null},{"author":"Ozzie_osman","children":[{"author":"c7b","children":[{"author":"Ozzie_osman","children":[{"author":"c7b","children":[],"created_at":"2026-10-10T13:42:26.000Z","created_at_i":1791639746,"id":50032892,"options":[],"parent_id":50032507,"points":null,"story_id":50023450,"text":"Sure, or let's say, no one else who was doing it before was noticed at this level. But hype doesn't automaically equal a paradigm shift. In this case, it's going back to what used to be one of the most common ML applications until not so long ago. If anything it feels like going old school (and I really mean that in a good way).","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T13:00:10.000Z","created_at_i":1791637210,"id":50032507,"options":[],"parent_id":50030518,"points":null,"story_id":50023450,"text":"I mean, it's obvious in retrospect, and technically not too hard to implement... Combining the power of LLMs with the constrained nature of classifiers. But no one did it well, and they did it first, and it took off.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T07:20:08.000Z","created_at_i":1791616808,"id":50030518,"options":[],"parent_id":50030503,"points":null,"story_id":50023450,"text":"Paradigm-shifting? A classifier? It's great and it's a cool use of Transformers, but AI really didn't only start with LLMs. We used to have an own term for AI models that can handle text instead of just numbers, now a model that outputs numbers instead of text is considered paradigm-shifting.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T07:16:51.000Z","created_at_i":1791616611,"id":50030503,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"It's possible that Jev has more in the pipeline. I mean, for shipping something so paradigm-shifting, this would be a bet on the founder/team continuing to do things faster or better.","title":null,"type":"comment","url":null},{"author":"fennecbutt","children":[{"author":"vrganj","children":[],"created_at":"2026-10-10T09:23:24.000Z","created_at_i":1791624204,"id":50031125,"options":[],"parent_id":50031096,"points":null,"story_id":50023450,"text":"Marx called this fictitious capital.
> Fictitious capital could be defined as a capitalisation on property ownership. Such ownership is real and legally enforced, as are the profits made from it, but the capital involved is fictitious; it is "money that is thrown into circulation as capital without any material basis in commodities or productive activivity".
https://en.wikipedia.org/wiki/Fictitious_capital","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T09:17:03.000Z","created_at_i":1791623823,"id":50031096,"options":[],"parent_id":50024954,"points":null,"story_id":50023450,"text":"Because the American stock market is all overvalued stuff with monopoly money.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:41:42.000Z","created_at_i":1791571302,"id":50024954,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"It's weird because two days after Jev was released there were a dozen decision models, a week later there are several dozen, mostly open source, OpenAI's own Decisions API [1] beats it, and you can easily finetune your own [2]. But as others have pointed out, this doesn't matter.
EDIT: As I wrote this Microsoft just released their own Decision-1 model [3].
[1] https://developers.openai.com/api/docs/guides/decisions
[2] https://unsloth.ai/docs/basics/train-your-own-decision-model...
[3] https://commandline.microsoft.com/microsoft-decision-1-model...","title":null,"type":"comment","url":null},{"author":"robotswantdata","children":[],"created_at":"2026-10-09T18:42:57.000Z","created_at_i":1791571377,"id":50024977,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"I\u2019ll just use open source, thanks","title":null,"type":"comment","url":null},{"author":"jypepin","children":[{"author":"CompleteSkeptic","children":[{"author":"stephen_cagle","children":[],"created_at":"2026-10-09T19:37:02.000Z","created_at_i":1791574622,"id":50025626,"options":[],"parent_id":50025424,"points":null,"story_id":50023450,"text":"Is it a pun or something? I have been called dense.
I actually resized my browser thinking maybe something weird was going on with flex-wrap or overflow or whatever it is.","title":null,"type":"comment","url":null},{"author":"aidos","children":[],"created_at":"2026-10-09T21:37:14.000Z","created_at_i":1791581834,"id":50026858,"options":[],"parent_id":50025424,"points":null,"story_id":50023450,"text":"With everything you must have going on right now, I love that you\u2019ve come on to explain your headline is an intentional joke. All sorts being thrown around in this thread, but this was definitely the important thing to clarify.
For what it\u2019s worth, however it works out, my guess is that the primitive Jev provides is likely to be considered essential in the future development of software.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:20:57.000Z","created_at_i":1791573657,"id":50025424,"options":[],"parent_id":50025020,"points":null,"story_id":50023450,"text":"intentional to make it fun (:","title":null,"type":"comment","url":null},{"author":"throw03172019","children":[{"author":"ambicapter","children":[],"created_at":"2026-10-09T19:38:16.000Z","created_at_i":1791574696,"id":50025638,"options":[],"parent_id":50025445,"points":null,"story_id":50023450,"text":"Jokes all around: https://jobs.ashbyhq.com/typesafe-ai/9a94651c-5d63-4e82-8854...","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:22:05.000Z","created_at_i":1791573725,"id":50025445,"options":[],"parent_id":50025020,"points":null,"story_id":50023450,"text":"They also say they are a fun team. Maybe it\u2019s a pun.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:46:33.000Z","created_at_i":1791571593,"id":50025020,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Their headline says "TypeSafe A raises series AI". Is that AI slope?","title":null,"type":"comment","url":null},{"author":"dovin","children":[{"author":"pickle-wizard","children":[],"created_at":"2026-10-09T19:16:34.000Z","created_at_i":1791573394,"id":50025374,"options":[],"parent_id":50025059,"points":null,"story_id":50023450,"text":"I just started experimenting with the decision models. I spun up Laya on a VM with a couple of vCPU and 6GB of RAM. I get the results in about half a second. No need for GPUs or tons of memory.
I am integrating it into the product I am building and to me it doesn't seem like there is much need to go with a SaaS for this since the requirements are so light. I just can run it in Cloud Run and get all of the scale I'll ever need, and I get to tell my customers their data never leaves my environment.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T18:49:30.000Z","created_at_i":1791571770,"id":50025059,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Jev does seem to have become the Kleenex of decision models. Is brand recognition worth $7.5B? There are lots of other decision models out there that perform at or near jev-level (laya, gliner 2.5 decide, even embedding gemma 2) that you can also run locally, and honestly I think this kind of model makes the most sense running locally as well. Maybe if TypeSafe can ship fast they can stay the default. Guess we'll find out.","title":null,"type":"comment","url":null},{"author":"jacobgold","children":[],"created_at":"2026-10-09T19:00:31.000Z","created_at_i":1791572431,"id":50025181,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Whether they can compete on decision models or not, TypeSafe showed that a lot of the market had missed something important. With this much money, they have a lot more chances to discover other important things that are missing.","title":null,"type":"comment","url":null},{"author":"christina97","children":[{"author":"soleveloper","children":[{"author":"tietjens","children":[{"author":"baobabKoodaa","children":[{"author":"tpetry","children":[{"author":"baobabKoodaa","children":[],"created_at":"2026-10-09T21:04:44.000Z","created_at_i":1791579884,"id":50026533,"options":[],"parent_id":50026282,"points":null,"story_id":50023450,"text":"The top alternative right now lists speed as faster than Jev?","title":null,"type":"comment","url":null},{"author":"senordevnyc","children":[],"created_at":"2026-10-09T23:24:32.000Z","created_at_i":1791588272,"id":50027825,"options":[],"parent_id":50026282,"points":null,"story_id":50023450,"text":"The space of problems that require super low latency AND are inputting so many tens of billions of tokens that the price matters, is a tiny one.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:38:02.000Z","created_at_i":1791578282,"id":50026282,"options":[],"parent_id":50025393,"points":null,"story_id":50023450,"text":"And the 'better' ones are slower and cost more for a tiny bit more accuracy. Its hard to sell that as being better when speed and price have been JEVs main selling points.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:18:01.000Z","created_at_i":1791573481,"id":50025393,"options":[],"parent_id":50025343,"points":null,"story_id":50023450,"text":"I'm not the person you're asking, but:
https://benchmarkheaven.com/jev-models
According to this benchmark, Jev is currently trailing Quyet-1.0-Large and a few other hastily put-together LLM-based decision API-like setups.","title":null,"type":"comment","url":null},{"author":"deepsquirrelnet","children":[],"created_at":"2026-10-09T19:21:33.000Z","created_at_i":1791573693,"id":50025434,"options":[],"parent_id":50025343,"points":null,"story_id":50023450,"text":"Here's another leaderboard:\nhttps://huggingface.co/spaces/multimodalart/jev-decision-ind...","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:14:08.000Z","created_at_i":1791573248,"id":50025343,"options":[],"parent_id":50025323,"points":null,"story_id":50023450,"text":"Who they trailing on quality?","title":null,"type":"comment","url":null},{"author":"nwhnwh","children":[],"created_at":"2026-10-09T19:54:11.000Z","created_at_i":1791575651,"id":50025813,"options":[],"parent_id":50025323,"points":null,"story_id":50023450,"text":"What even is this? What does it do?","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:12:21.000Z","created_at_i":1791573141,"id":50025323,"options":[],"parent_id":50025232,"points":null,"story_id":50023450,"text":"They don't lead on latency nor quality; but their execution was superb","title":null,"type":"comment","url":null},{"author":"verdverm","children":[],"created_at":"2026-10-09T19:24:05.000Z","created_at_i":1791573845,"id":50025469,"options":[],"parent_id":50025232,"points":null,"story_id":50023450,"text":"Wonder if they can get coin flips and dice rolls to make sense with this fresh funding, or if it even matters to people.
I have no faith in the technique if it cannot do the basics (i.e. not real probabilities, the confidence for coin flip outcomes)
tried it a couple of days ago here: https://jevplayground.com
the "not real probability" disclaimer only appears after you get a result","title":null,"type":"comment","url":null},{"author":"rvz","children":[{"author":"objektif","children":[{"author":"JamesSwift","children":[],"created_at":"2026-10-10T02:18:26.000Z","created_at_i":1791598706,"id":50028881,"options":[],"parent_id":50026655,"points":null,"story_id":50023450,"text":"35% 10% 55%","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:17:03.000Z","created_at_i":1791580623,"id":50026655,"options":[],"parent_id":50025576,"points":null,"story_id":50023450,"text":"More like 10% 10% 80%.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:32:37.000Z","created_at_i":1791574357,"id":50025576,"options":[],"parent_id":50025232,"points":null,"story_id":50023450,"text":"This is all due to 40% marketing, 50% execution and 10% credentials (with the founders being associated with creating ChatGPT).
If anyone else came up with the same concept on a Reddit thread (they have) it no-one would care without those characteristics even if you are "first".
Rebranding, execution, marketing, ex-<big_name_company> and mostly importantly, hype is what gets the investors scrambling into throwing money at you.","title":null,"type":"comment","url":null},{"author":"binlog","children":[{"author":"redanddead","children":[],"created_at":"2026-10-09T21:11:06.000Z","created_at_i":1791580266,"id":50026603,"options":[],"parent_id":50026538,"points":null,"story_id":50023450,"text":"Every startup is priced as an acq target","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:05:29.000Z","created_at_i":1791579929,"id":50026538,"options":[],"parent_id":50025232,"points":null,"story_id":50023450,"text":"The model itself is a negligible part of the valuation. The company is priced as an acquisition target.","title":null,"type":"comment","url":null},{"author":"hbrn","children":[{"author":"vanuatu","children":[{"author":"hbrn","children":[{"author":"qlte","children":[{"author":"hbrn","children":[],"created_at":"2026-10-09T23:48:07.000Z","created_at_i":1791589687,"id":50027982,"options":[],"parent_id":50027748,"points":null,"story_id":50023450,"text":"Ha, two years is too generous.
If they weren't lying about $100M ARR, they had to be at 6.5T/day for a week to hit that number.
They also claimed they were at 1T/day just couple weeks ago. That's $15M ARR.
85% drop in just a couple of weeks?","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:12:29.000Z","created_at_i":1791587549,"id":50027748,"options":[],"parent_id":50027513,"points":null,"story_id":50023450,"text":"Especially for a company still in their 15 minutes of fame that could very plausibly turn out to be a one hit wonder and a trivia question in two years.
Multiplying out revenue from the peak of their mini hype cycle while a dozen well financed competitors target them directly seems extremely optimistic.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:43:59.000Z","created_at_i":1791585839,"id":50027513,"options":[],"parent_id":50027378,"points":null,"story_id":50023450,"text":"Measuring ARR based on 7 days of data is beyond stupid.
You can create a company with 2B shares and sell one share to your friend for $1000. Lo and behold, you own $2 trillion dollar company, leaving Elon behind.","title":null,"type":"comment","url":null},{"author":"dannyw","children":[{"author":"nikanj","children":[],"created_at":"2026-10-10T12:59:03.000Z","created_at_i":1791637143,"id":50032495,"options":[],"parent_id":50032307,"points":null,"story_id":50023450,"text":"It\u2019s more likely just \u201dMarketing put that claim on the page, who could ever disprove it?\u201d","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T12:33:46.000Z","created_at_i":1791635626,"id":50032307,"options":[],"parent_id":50027378,"points":null,"story_id":50023450,"text":"1/3 of the Fortune 500 is likely measured by something like a single @company.com put a credit card in, against their company\u2019s procurement and security policies.
It\u2019s not to be unexpected for a new product and startup, but they lack SOC 2 certification, and the other things enterprises look for on information security.
I\u2019d be surprised if many procurement divisions are approving typesafe AI, and I fear by the time they achieve it, existing vendors would\u2019ve catched up.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T22:29:48.000Z","created_at_i":1791584988,"id":50027378,"options":[],"parent_id":50026823,"points":null,"story_id":50023450,"text":"since their release they surpassed 100M arr and 1/3 of the F500
its clear they own the mindshare around this type of primitive which is a massive premium","title":null,"type":"comment","url":null},{"author":"CSDude","children":[{"author":"randysalami","children":[],"created_at":"2026-10-10T13:47:51.000Z","created_at_i":1791640071,"id":50032955,"options":[],"parent_id":50031402,"points":null,"story_id":50023450,"text":"Silly question but is it right? I am in grad school and this semester we run multimodal data through ML and LLM-as-judge classifiers. Even plugging in ChatGPT with a relevant judge prompt produces relatively believable results. At the same time, I struggle to trust it to generalize. I\u2019ve read that Jev does classification across many domains and is also powered by an LLM back-end? What separates it from my above approach outside of harness efforts and finetuning of the LLM? Thank you!","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T10:17:30.000Z","created_at_i":1791627450,"id":50031402,"options":[],"parent_id":50026823,"points":null,"story_id":50023450,"text":"I built a NAND gate with Jev to contribute to useless demos: https://x.com/mustafaakin/status/2103574428475154635
But economics of Jev enable classifying high volume things. And I've tried all the open-weight Qwen based decision models on my Spark and none of them came closer to quality and suck at batch inference so they are doing something custom.","title":null,"type":"comment","url":null},{"author":"baby","children":[],"created_at":"2026-10-10T12:38:57.000Z","created_at_i":1791635937,"id":50032351,"options":[],"parent_id":50026823,"points":null,"story_id":50023450,"text":"What? The moment this was released we integrated decision models EVERYWHERE. If you\u2019re not using it you\u2019re basically like these people who think LLMs are useless lol","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:34:39.000Z","created_at_i":1791581679,"id":50026823,"options":[],"parent_id":50025232,"points":null,"story_id":50023450,"text":"> people that came up with a product people wanted
We've yet to see whether this is true, or is it just manufactured demand. There are dozens of Jev demos, but pretty much all of them are either cool but useless, or simply fake (i.e. harness doing 99% of the work).","title":null,"type":"comment","url":null},{"author":"cootsnuck","children":[{"author":"asa123","children":[{"author":"adrianmsmith","children":[{"author":"asa123","children":[],"created_at":"2026-10-10T14:23:40.000Z","created_at_i":1791642220,"id":50033326,"options":[],"parent_id":50030637,"points":null,"story_id":50023450,"text":"i mean marketing in some sense \u201cfeels\u201d like fraud when presented in some very specific way, no? at some level it\u2019s: generate interest about a thing when there didn\u2019t exist any interest before. sometimes they generate interest about something that has no substance and that feels kind of fraudulent (this isn\u2019t that case)
but in this case, this is product extremely boosted for super viral marketing, and this is the first time i\u2019ve read about this boost since hearing of the product whereas everybody only really talks of its substance. feels like a magician kept a trick going for a very long time, people forgot they were watching a magic show, then the curtains come down and they reveal the deceit after (i\u2019m being slightly melodramatic)","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T07:41:32.000Z","created_at_i":1791618092,"id":50030637,"options":[],"parent_id":50029379,"points":null,"story_id":50023450,"text":"Why?
I read the attached links and it just looks like a marketing agency?","title":null,"type":"comment","url":null},{"author":"Den_VR","children":[],"created_at":"2026-10-10T10:58:41.000Z","created_at_i":1791629921,"id":50031659,"options":[],"parent_id":50029379,"points":null,"story_id":50023450,"text":"Why not both. The engineering side isn\u2019t fraud, but the business side is doing business things \u201cfor the sake of the project.\u201d","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T03:53:13.000Z","created_at_i":1791604393,"id":50029379,"options":[],"parent_id":50028445,"points":null,"story_id":50023450,"text":"you\u2019ve got to be kidding me. i know it ISNT fraud but i can\u2019t help but feel like this SMELLS like fraud","title":null,"type":"comment","url":null},{"author":"O_H_E","children":[],"created_at":"2026-10-10T14:25:41.000Z","created_at_i":1791642341,"id":50033344,"options":[],"parent_id":50028445,"points":null,"story_id":50023450,"text":"Oh wow, thanks for this.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T01:03:04.000Z","created_at_i":1791594184,"id":50028445,"options":[],"parent_id":50025232,"points":null,"story_id":50023450,"text":"> On top of that they have very strong marketing muscle that took the AI world by storm.
Not them per se, but doomers.ai [0] which is a uh... "launch virality agency".
So less that Typesafe has a strong marketing muscle, and more that they paid at least $100K [1] for "organic" buzz.
[0] - https://doomers.ai/work/typesafe-ai-case-study
[1] - https://doomers.ai/guides/ai-marketing-agency","title":null,"type":"comment","url":null},{"author":"solatic","children":[],"created_at":"2026-10-10T04:52:36.000Z","created_at_i":1791607956,"id":50029686,"options":[],"parent_id":50025232,"points":null,"story_id":50023450,"text":"A great team is now worth $7.5B?
They can be both great and over-valued at the same time.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:05:18.000Z","created_at_i":1791572718,"id":50025232,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Everyone appears surprised by this news. It\u2019s clear that they don\u2019t have a product with some incredible moat. But they clearly have good engineering and product people that came up with a product people wanted. On top of that they have very strong marketing muscle that took the AI world by storm. And as far as I\u2019ve seen, they still lead in some part of the latency-quality (-cost) curve?
They may well be a good team to throw money behind if you are hoping to bet on a new AI lab.","title":null,"type":"comment","url":null},{"author":"maxdo","children":[{"author":"kylehotchkiss","children":[],"created_at":"2026-10-09T21:15:33.000Z","created_at_i":1791580533,"id":50026643,"options":[],"parent_id":50025334,"points":null,"story_id":50023450,"text":"See clef","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:13:23.000Z","created_at_i":1791573203,"id":50025334,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"interestingly , it destroy the landscape of Chinese models.
Unless china takes leadership in frontier space the picture is next :
1. cheap workhorses for classification, routing, other scenarios : Jev\n2. coding agents with less erros : Anthropic/Openai, etc. \n3. Science /Legal/Medical : A mixture of Jev+Anthropic scenarios","title":null,"type":"comment","url":null},{"author":"dvt","children":[{"author":"minimaxir","children":[],"created_at":"2026-10-09T19:21:02.000Z","created_at_i":1791573662,"id":50025426,"options":[],"parent_id":50025377,"points":null,"story_id":50023450,"text":"There have been a lot of posts/comments claiming "Jev-like models" but that's more of an shorthand for decision models, not astroturfing.","title":null,"type":"comment","url":null},{"author":"denverllc","children":[{"author":"amarcheschi","children":[{"author":"StrLght","children":[{"author":"amarcheschi","children":[],"created_at":"2026-10-10T12:15:22.000Z","created_at_i":1791634522,"id":50032163,"options":[],"parent_id":50031898,"points":null,"story_id":50023450,"text":"Oh this is interesting. I've seen it was added in the last months on x but it's cool nonetheless. Otoh, i don't miss not using a lot of social networks except for a few niche subreddits","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T11:39:59.000Z","created_at_i":1791632399,"id":50031898,"options":[],"parent_id":50031789,"points":null,"story_id":50023450,"text":"It's mentioned in video, here's an excerpt from FAQ [0]:
> Do creators disclose that a post is paid?
> Yes, every time. Every creator post we place carries X's paid partnership label. That is an FTC requirement for any post where a brand has paid or given something of value, and it is X's own rule for sponsored content.
Although, there are other services, that sound shady.
[0]: https://doomers.ai/faq","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T11:20:34.000Z","created_at_i":1791631234,"id":50031789,"options":[],"parent_id":50026297,"points":null,"story_id":50023450,"text":"Just a question, are these creators paid to sponsor these products? Because at least in eu, they would need to add a "#adv" in their post, which makes it easy for me to spot sponsored content here. I guess in the rest of the world it may not be necessary","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:39:07.000Z","created_at_i":1791578347,"id":50026297,"options":[],"parent_id":50025377,"points":null,"story_id":50023450,"text":"Yes, it's astroturfed everywhere (like X and reddit).
https://www.youtube.com/watch?v=xNgQtzEl4lY
Jev is used as an example of a successful marketing launch where they worked with many X "creators" prior to its release, so that all the creators would repost to put it to the top of everyone's feed. Then, over the following days they'd repost so it maintained momentum.
See: doomers.ai, clickstrike, growth matrix, etc. They use coordinated engagement, paid influencer networks, customized messaging, etc.
Jev isn't a terrible product, but it's way overhyped.","title":null,"type":"comment","url":null},{"author":"softwaredoug","children":[{"author":"dvt","children":[],"created_at":"2026-10-09T22:41:10.000Z","created_at_i":1791585670,"id":50027491,"options":[],"parent_id":50026713,"points":null,"story_id":50023450,"text":"Because Jev is basically a "generalized classifier" which.. doesn't quite make much sense. Training a classifier is pretty easy and has been done routinely for like two decades now. Classifiers also tend to be very localized; for example, I've worked on classifiers that would bucket web traffic into "potential buyer" or "potential seller"\u2014but this was very specific to the use case (vehicles, in our case).
If I seriously needed a classifier, I would just train my own and it would run on an iPhone. Any CTO worth their salt would suggest the same, because it's not even remotely comparable to training a large language model (w.r.t. compute or training data required).","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T21:21:57.000Z","created_at_i":1791580917,"id":50026713,"options":[],"parent_id":50025377,"points":null,"story_id":50023450,"text":"Why is it a middling product? Most clones don\u2019t approach its performance, and it solves a specific problem well in a way that was awkward and ignored by most frontier labs.","title":null,"type":"comment","url":null},{"author":"vanuatu","children":[],"created_at":"2026-10-09T22:32:09.000Z","created_at_i":1791585129,"id":50027405,"options":[],"parent_id":50025377,"points":null,"story_id":50023450,"text":"distribution is a moat","title":null,"type":"comment","url":null},{"author":"incompressible","children":[{"author":"dvt","children":[{"author":"hall0ween","children":[],"created_at":"2026-10-10T04:04:02.000Z","created_at_i":1791605042,"id":50029449,"options":[],"parent_id":50028694,"points":null,"story_id":50023450,"text":"I appreciate you putting words to the astro-turfing. two areas llms have helped me are in (i) giving me work and (ii) learning how people in my life repeat what they hear from \u201cknowledgeable\u201d people. And not saying the voices proselytizing llms aren\u2019t knowledgeable but I bet there\u2019s an agenda behind their words (ie \u201ci have something powerful\u201d and \u201cgive me money!)","title":null,"type":"comment","url":null},{"author":"kanzure","children":[],"created_at":"2026-10-10T12:55:13.000Z","created_at_i":1791636913,"id":50032460,"options":[],"parent_id":50028694,"points":null,"story_id":50023450,"text":"> Kind of sad seeing so many \u201ctech people\u201d (including here on HN) falling for it hook, line, and sinker.
Unfortunately it is not sufficient for someone to think "I am immune to marketing" to actually be immune against marketing.","title":null,"type":"comment","url":null},{"author":"Analemma_","children":[],"created_at":"2026-10-10T16:01:49.000Z","created_at_i":1791648109,"id":50034226,"options":[],"parent_id":50028694,"points":null,"story_id":50023450,"text":""People who think they're immune to marketing are the biggest suckers of all" always been bad, but it's worse than ever because a bunch of people have decided that the most reliable way to judge AI performance is its vibes on X The Everything App. I can't even really fault places like doomers.ai when their marks are practically asking to get conned.","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T01:46:37.000Z","created_at_i":1791596797,"id":50028694,"options":[],"parent_id":50028673,"points":null,"story_id":50023450,"text":"Cool, yeah my spidey sense was definitely going off. It\u2019s a half-decent idea, but it does appear that marketing is all you need.
Kind of sad seeing so many \u201ctech people\u201d (including here on HN) falling for it hook, line, and sinker.","title":null,"type":"comment","url":null},{"author":"amarcheschi","children":[],"created_at":"2026-10-10T11:22:39.000Z","created_at_i":1791631359,"id":50031801,"options":[],"parent_id":50028673,"points":null,"story_id":50023450,"text":"Jesus f christ that might be why I saw it for a few days here on hn together with collateral discussions and topics","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T01:42:18.000Z","created_at_i":1791596538,"id":50028673,"options":[],"parent_id":50025377,"points":null,"story_id":50023450,"text":"Just found out from a previous comment: https://doomers.ai/work/typesafe-ai-case-study
Obscene marketing is all you need it seems.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:16:54.000Z","created_at_i":1791573414,"id":50025377,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Is Jev being astroturfed on HN? It certainly feels like it. It's a middling product with virtually no moat (but great marketing).","title":null,"type":"comment","url":null},{"author":"amelius","children":[{"author":"benatkin","children":[],"created_at":"2026-10-09T19:39:08.000Z","created_at_i":1791574748,"id":50025646,"options":[],"parent_id":50025433,"points":null,"story_id":50023450,"text":"They weren't running on OpenAI, so nope.","title":null,"type":"comment","url":null},{"author":"vilos1611","children":[{"author":"denverllc","children":[],"created_at":"2026-10-09T20:43:29.000Z","created_at_i":1791578609,"id":50026354,"options":[],"parent_id":50026075,"points":null,"story_id":50023450,"text":"> but I question how much folks are actually experimenting with what's out there
I did. Originally I had a project that I had been wanting to do and thought to use a decision model for it. Jev, OpenAI, etc. are all within percentage points of each other.
Then I used traditional ML and found a small classifier (gemma 4) with traditional embeddings worked 2x as well.
Jev is the general purpose ML pipeline for when you want average results. Nearly every application has a "better" option available with a small amount of work.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:18:23.000Z","created_at_i":1791577103,"id":50026075,"options":[],"parent_id":50025433,"points":null,"story_id":50023450,"text":"We've tested the decisions API against Jev at work and it's worse in various dimensions. Costs more, higher error rates, slower, and the answers are worse.
A lot of people are shouting about how Jev hasn't actually differentiated itself, but I question how much folks are actually experimenting with what's out there before coming up with an opinion.
For us, it's cleae that OpenAI rushed this out to meet the hype in the market right now without having a product that actually meets the bar Jev has set.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:21:31.000Z","created_at_i":1791573691,"id":50025433,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"They already got Sherlocked by OpenAI:
https://developers.openai.com/api/docs/guides/decisions","title":null,"type":"comment","url":null},{"author":"stickfigure","children":[{"author":"esafak","children":[{"author":"shimman","children":[],"created_at":"2026-10-10T00:58:22.000Z","created_at_i":1791593902,"id":50028403,"options":[],"parent_id":50026352,"points":null,"story_id":50023450,"text":"With any luck it'll radicalize a new generation of workers to rise up and stop SV.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:43:22.000Z","created_at_i":1791578602,"id":50026352,"options":[],"parent_id":50025438,"points":null,"story_id":50023450,"text":"The needle has moved. This is a wonderful thing.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:21:47.000Z","created_at_i":1791573707,"id":50025438,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Remember when reaching $1B valuation made you an exotic "unicorn"?","title":null,"type":"comment","url":null},{"author":"enahs-sf","children":[],"created_at":"2026-10-09T19:29:29.000Z","created_at_i":1791574169,"id":50025537,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"got the recruiter call only to essentially be summarily rejected because my pedigree is wack. looks like i would've gotten hosed on valuation anyways.","title":null,"type":"comment","url":null},{"author":"maherbeg","children":[{"author":"santiago-pl","children":[{"author":"TN1ck","children":[{"author":"maherbeg","children":[],"created_at":"2026-10-09T20:42:08.000Z","created_at_i":1791578528,"id":50026339,"options":[],"parent_id":50026127,"points":null,"story_id":50023450,"text":"This is awesome! Please submit updates when a new model comes out that people are saying are better than Jev!","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:22:49.000Z","created_at_i":1791577369,"id":50026127,"options":[],"parent_id":50025807,"points":null,"story_id":50023450,"text":"I also benchmarked them here [1] for content moderation. Jev is better than any of them.
[1]: https://tn1ck.com/blog/jevdit","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:53:40.000Z","created_at_i":1791575620,"id":50025807,"options":[],"parent_id":50025746,"points":null,"story_id":50023450,"text":"I found this benchmark helpful: https://benchmarkheaven.com/jev-models","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T19:48:44.000Z","created_at_i":1791575324,"id":50025746,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Has anyone actually eval'd the other open source options against Jev on real world tasks rather than looking at benchmarks?
I see a lot of people parroting the quick open source alternatives as being better on the benchmarks, but it's such a new category that I'm not convinced we have solid benchmarks.
I'm hoping a company releases an internal eval benchmark for these options. I'm sure some of the open source ones are solid in some cases, but would love to see more reliable data.","title":null,"type":"comment","url":null},{"author":"mococa","children":[],"created_at":"2026-10-09T19:50:22.000Z","created_at_i":1791575422,"id":50025770,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"~7B for a thing that we already have open-source?","title":null,"type":"comment","url":null},{"author":"rokhayakebe","children":[{"author":"fnoef","children":[],"created_at":"2026-10-10T08:58:17.000Z","created_at_i":1791622697,"id":50031003,"options":[],"parent_id":50026068,"points":null,"story_id":50023450,"text":"Koenigsegg for every founding member?","title":null,"type":"comment","url":null},{"author":"ac29","children":[],"created_at":"2026-10-10T12:57:23.000Z","created_at_i":1791637043,"id":50032481,"options":[],"parent_id":50026068,"points":null,"story_id":50023450,"text":"Compute and payroll","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:17:38.000Z","created_at_i":1791577058,"id":50026068,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Can someone who actually knows these things share how might a company like this spend $870M over the years?","title":null,"type":"comment","url":null},{"author":"pseudotensor","children":[{"author":"pseudotensor","children":[],"created_at":"2026-10-09T21:45:12.000Z","created_at_i":1791582312,"id":50026966,"options":[],"parent_id":50026118,"points":null,"story_id":50023450,"text":"Update: Microsoft just announced Microsoft-Decision-1 (a post-trained Qwen3.5-9B) and benchmarked it against H2O-Lightning-4B, among others: https://commandline.microsoft.com/microsoft-decision-1-model...
One note on their speed comparison: the JevBench board shows self-hosted models with an adjusted latency of "2x + 0.15 s (assumption, not measured)". Our measured p50 is 29 ms; the adjusted figure is 0.21 s. They report 85 ms for theirs.","title":null,"type":"comment","url":null},{"author":"smrtinsert","children":[{"author":"pseudotensor","children":[],"created_at":"2026-10-10T02:09:12.000Z","created_at_i":1791598152,"id":50028826,"options":[],"parent_id":50027655,"points":null,"story_id":50023450,"text":"Will do!","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T23:01:24.000Z","created_at_i":1791586884,"id":50027655,"options":[],"parent_id":50026118,"points":null,"story_id":50023450,"text":"Is it possible to quantize the model? Don't think I want to run 8 gigs on my local just for decisions but curious about it for sure","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T20:22:08.000Z","created_at_i":1791577328,"id":50026118,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Disclosure: I work at H2O.ai.
We released an Apache-2.0, open-weight 4B decision model that scores above Jev 1.13 on JevBench's composite score (72.5 vs 71.5) and is currently the top open model there: https://benchmarkheaven.com/jev-models . Newer models coming even larger than beat Jev in intelligence as well.
- Same contract as Jev: state + typed questions in, calibrated probabilities out, one forward pass, no generated tokens.\n- Your data never leaves your environment, and there's no per-call fee.
Weights, card and run instructions: https://huggingface.co/h2oai/h2o-lightning-4b","title":null,"type":"comment","url":null},{"author":"collimarco","children":[],"created_at":"2026-10-09T20:27:56.000Z","created_at_i":1791577676,"id":50026176,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"This looks like the top of the dot-com bubble...","title":null,"type":"comment","url":null},{"author":"axus","children":[],"created_at":"2026-10-09T20:29:01.000Z","created_at_i":1791577741,"id":50026190,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"So... are they worth more, or less than 0xide","title":null,"type":"comment","url":null},{"author":"woadwarrior01","children":[],"created_at":"2026-10-09T20:31:02.000Z","created_at_i":1791577862,"id":50026214,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"I wonder how much of it has been earmarked for astroturfing on HN and X? :D","title":null,"type":"comment","url":null},{"author":"kevinkatzke","children":[],"created_at":"2026-10-09T20:31:21.000Z","created_at_i":1791577881,"id":50026218,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"A few year ago you could IPO at this valuation.","title":null,"type":"comment","url":null},{"author":"TeMPOraL","children":[],"created_at":"2026-10-09T20:31:34.000Z","created_at_i":1791577894,"id":50026220,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Only just now I realized that "TypeSafe AI" are the people behind Jev, and "System One" isn't the company name, as I assumed, but a larger project label.","title":null,"type":"comment","url":null},{"author":"Thorentis","children":[],"created_at":"2026-10-09T20:35:48.000Z","created_at_i":1791578148,"id":50026261,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Another data point confirming that AI is a bubble.","title":null,"type":"comment","url":null},{"author":"sajithdilshan","children":[],"created_at":"2026-10-09T20:44:49.000Z","created_at_i":1791578689,"id":50026363,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Is there a second wave of AI bubble happening? How can an AI classifier company be worth of $7.5B","title":null,"type":"comment","url":null},{"author":"softwaredoug","children":[],"created_at":"2026-10-09T21:19:00.000Z","created_at_i":1791580740,"id":50026675,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"What\u2019s interesting about the Jev moment isn\u2019t just Jev, it\u2019s the unleashing of distillation / fine tuning outside the frontier-adjacent labs. It\u2019s the sudden explosion of a million Jevs.
If being an \u201cAI Researcher\u201d is a ticket to multimillion dollar salary, AI training talent cannot be contained to a handful of companies. It\u2019ll become more common and diffuse. The old advice of not fine tuning, because it\u2019s hard, goes out the window as that knowledge diffuses through the industry.
A similar thing is happening in search. For a long time labs have trained tailored embedding models. And now companies like SID training their own agentic models that are smaller and faster at search than GPT-5.","title":null,"type":"comment","url":null},{"author":"gdiamos","children":[],"created_at":"2026-10-09T21:42:24.000Z","created_at_i":1791582144,"id":50026939,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Good job typesafe.
You won the competition with VCs","title":null,"type":"comment","url":null},{"author":"atleastoptimal","children":[],"created_at":"2026-10-09T21:45:51.000Z","created_at_i":1791582351,"id":50026976,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"I think this is a hedge against major AI regulation.
Invariably near-AGI systems created by OpenAI/Anthropic will be very destabalizing. In the end the world will probably regulate AI capable of [any] <-> [any] input/output types. Models will need to be limited on their outputs by law so they cannot have unbounded, unpredictable outcomes. Jev is the ideal version of "benefits of AI without making humans obsolete" that might be the consensus once the track superhuman AI and its consequences are clear.","title":null,"type":"comment","url":null},{"author":"dmix","children":[],"created_at":"2026-10-09T22:03:44.000Z","created_at_i":1791583424,"id":50027149,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"That press release made me cringe a bit. Maybe corporate speak wasn't so bad after all.","title":null,"type":"comment","url":null},{"author":"minimaxa","children":[],"created_at":"2026-10-09T22:43:01.000Z","created_at_i":1791585781,"id":50027505,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"No one finds it strange that every defector from Opensi gets half a billion dollars in funding and already has a third of Fortune 500 companies paying for their product in a few weeks with a product that's already been replicated 10 times very easily?
OK, I'm jealous... Lol","title":null,"type":"comment","url":null},{"author":"greesil","children":[],"created_at":"2026-10-10T01:51:20.000Z","created_at_i":1791597080,"id":50028718,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"The amount of investment being thrown at this when we have serious existential problems as a species is how I know the system is definitively broken. Capitalism got us this far, but it's time for something new.","title":null,"type":"comment","url":null},{"author":"ada1981","children":[],"created_at":"2026-10-10T02:52:35.000Z","created_at_i":1791600755,"id":50029061,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"I built JEV-search and JEV-chat and have been really happy with the results\u2026
Curious to see how all of these can work together, and giving LLMs the ability to deploy their own workflows and built type 1 systems as they go has also been fun and useful.","title":null,"type":"comment","url":null},{"author":"nico","children":[],"created_at":"2026-10-10T05:00:47.000Z","created_at_i":1791608447,"id":50029725,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Hopefully some of the Typesafe hype spills over this way to Jeffy[0]
Open source[1]. It comes with 68 pre trained classifiers which run and train on CPU alone. They run locally, are faster than Jev/Laya/Decisions, and they can perform many different tasks; from labeling email, all the way to playing Doom[2]
[0] https://jeffyclassify.com/
[1] https://github.com/nicobrenner/jeffy
[2] https://playground.jeffyclassify.com/#doom","title":null,"type":"comment","url":null},{"author":"jdw64","children":[],"created_at":"2026-10-10T05:24:03.000Z","created_at_i":1791609843,"id":50029843,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"I only use Jev at the tool-calling layer to keep tool costs low, but I don't really see what the big advantage is.","title":null,"type":"comment","url":null},{"author":"villgax","children":[],"created_at":"2026-10-10T06:52:37.000Z","created_at_i":1791615157,"id":50030395,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Lol","title":null,"type":"comment","url":null},{"author":"2001zhaozhao","children":[],"created_at":"2026-10-10T06:54:54.000Z","created_at_i":1791615294,"id":50030403,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Tbf this is THE way to get cost-effective AI into video games (think enemy AIs and NPC AIs that are much better than current ones and runnable on consumer GPUs), i'm really looking forward for decision models maturing.","title":null,"type":"comment","url":null},{"author":"alpineman","children":[],"created_at":"2026-10-10T07:05:39.000Z","created_at_i":1791615939,"id":50030448,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"From Andreessen the Trump donor. Sad, will switch to something else now.","title":null,"type":"comment","url":null},{"author":"_davide_","children":[],"created_at":"2026-10-10T07:35:45.000Z","created_at_i":1791617745,"id":50030606,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"I guess they are investing in the team rather than on their demo product","title":null,"type":"comment","url":null},{"author":"dj_io","children":[],"created_at":"2026-10-10T07:40:18.000Z","created_at_i":1791618018,"id":50030630,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"There is significant congregation of investment towards companies like this whereas many other with meaningful innovation are struggling to survive. The model is not a revolution like I must say how llm did at the face value. Yes, transformers were there for long but the use case etc. If you do cost comparison using regular model and setting it with prompt and structured output you can get result very close. This is what most open Jev and other are doing with tuning llm.","title":null,"type":"comment","url":null},{"author":"Culonavirus","children":[],"created_at":"2026-10-10T08:37:50.000Z","created_at_i":1791621470,"id":50030923,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Oracle needs to hurry up and get downgraded to junk status at which point it will raise shit and at which point it will finally send off the avalanche of normalization across this industry.
There needs to be cleansing with fire. Weeds need to die. Trees need their branches cut. The sooner the better.
A shitty ass random "ai" startup built entirely on hype and astroturfing should not be raising anywhere near this amount of money.","title":null,"type":"comment","url":null},{"author":"xxmarkuski","children":[],"created_at":"2026-10-10T08:39:51.000Z","created_at_i":1791621591,"id":50030932,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Their job opening as Safety Javeroni was a good laugh too [0]
[0] https://jobs.ashbyhq.com/typesafe-ai/9a94651c-5d63-4e82-8854...","title":null,"type":"comment","url":null},{"author":"pythonRon","children":[],"created_at":"2026-10-10T09:42:52.000Z","created_at_i":1791625372,"id":50031226,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"If Typesafe really wants to do some good in the world, it'll raise money for those in need.","title":null,"type":"comment","url":null},{"author":"noduerme","children":[],"created_at":"2026-10-10T10:46:50.000Z","created_at_i":1791629210,"id":50031578,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"If you ask the right question, you don't need a decision model.
If you don't ask the right question, you might as well consult a psychic.
If all you need is scale, I guess decision models are cheaper than psychics.","title":null,"type":"comment","url":null},{"author":"thatsadude","children":[],"created_at":"2026-10-10T10:57:10.000Z","created_at_i":1791629830,"id":50031652,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"These investors are not that smart","title":null,"type":"comment","url":null},{"author":"999900000999","children":[],"created_at":"2026-10-10T11:40:09.000Z","created_at_i":1791632409,"id":50031899,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"What a neat API.
I went ahead and applied for a job, and unlike Oxide, TypeSafe has an application that only takes a few minutes to fill out.
I did take a second to play with the TypeSafe API first. I'm also excited to get Jev running locally.
Not like I'm realistically good enough for a top AI company, but I can dream. I got multi code signal tests for Anthropic and got 70% at best.
Would be nice to get equity, ipo and cash out.
I don't want to work in my 50s","title":null,"type":"comment","url":null},{"author":"fwlr","children":[],"created_at":"2026-10-10T11:40:59.000Z","created_at_i":1791632459,"id":50031908,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Others have capably pointed out why their technology cannot be much of this valuation. I submit that \u201cthe ability to grab attention in the AI room\u201d is what is being valued this highly.","title":null,"type":"comment","url":null},{"author":"Amekedl","children":[],"created_at":"2026-10-10T12:03:09.000Z","created_at_i":1791633789,"id":50032061,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"This makes no sense, unless they are going to crack something, like the actual atrocious accuracy of around 70%.
It's always been the last 30% where it just fails.
Guess it is 870M to 0","title":null,"type":"comment","url":null},{"author":"-HIMAO-","children":[{"author":"sixhobbits","children":[],"created_at":"2026-10-10T12:51:33.000Z","created_at_i":1791636693,"id":50032442,"options":[],"parent_id":50032335,"points":null,"story_id":50023450,"text":"Definitely intentional, they're fun","title":null,"type":"comment","url":null}],"created_at":"2026-10-10T12:36:40.000Z","created_at_i":1791635800,"id":50032335,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"This is pretty petty but they got the title wrong (should've been 'Typesafe AI Raises Series A' not 'Typesafe A Raises Series AI')","title":null,"type":"comment","url":null},{"author":"felix089","children":[],"created_at":"2026-10-10T13:44:33.000Z","created_at_i":1791639873,"id":50032919,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Good timing with our pac-man benchmark for decision models, now also including Microsoft Decision-1: https://news.ycombinator.com/item?id=50007993","title":null,"type":"comment","url":null},{"author":"r4indeer","children":[],"created_at":"2026-10-10T14:02:08.000Z","created_at_i":1791640928,"id":50033107,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Not a native English speaker; am I missing some kind of wordplay here or is the "I" in the wrong place in the title? It says: "TypeSafe A raises Series AI", but shouldn't it be "TypeSafe AI raises Series A"?","title":null,"type":"comment","url":null},{"author":"pu_pe","children":[],"created_at":"2026-10-10T15:38:16.000Z","created_at_i":1791646696,"id":50033987,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"This might be the pets.com moment of this era.","title":null,"type":"comment","url":null},{"author":"stymaar","children":[],"created_at":"2026-10-10T16:00:32.000Z","created_at_i":1791648032,"id":50034214,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"Multi-hundred million raised at a multi-billion valuation for a series A is literally insane.
How are the investors hoping to get their money back? With a $20B acqui-hire from X.ai later this year or what's the plan?","title":null,"type":"comment","url":null},{"author":"fHr","children":[],"created_at":"2026-10-10T16:36:23.000Z","created_at_i":1791650183,"id":50034545,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"this market lol","title":null,"type":"comment","url":null},{"author":"0x10ca1h0st","children":[],"created_at":"2026-10-11T05:45:10.000Z","created_at_i":1791697510,"id":50040136,"options":[],"parent_id":50023450,"points":null,"story_id":50023450,"text":"It's funny how Andreessen Horowitz will go on a media blitz to ensure any company they invest in gets priority while Laya gets ignored. Hilarious.","title":null,"type":"comment","url":null}],"created_at":"2026-10-09T17:02:31.000Z","created_at_i":1791565351,"id":50023450,"options":[],"parent_id":null,"points":444,"story_id":50023450,"text":null,"title":"Typesafe AI raises $870M at $7.5B","type":"story","url":"https://typesafe.ai/blog/series-ai"}