Rough summary of the available public information:
- They had an accident in 2023 injuring, by not stopping instantly, a passerby who was hit by another car.
- They were investigated and then California temporarily suspended their license to drive around.
- By the end of 2023 there was a massive layoff, including top executives leaving. There were claims that the operation was too human-dependent / required too much human interventions.
- In 2024 there were some news that they were starting up again (outside CA maybe).
- In Dec 2024, GM dropped their funding altogether and gave up on self-driving / robotaxi, refocusing on driver assistance instead. Some analysts claimed that was in part (in addition to finances) due to politics and them not wanting to face off with Tesla under the new administration.
It always seemed to me that something was missing from the overall picture:
- While second to Waymo, they still seemed already far ahead from others; maybe they were, say, 90% there.
- Others, like Tesla, had many accidents, including fatal ones. But for those the gauntlet was not thrown that quickly. It is expected that self-driving cars will have accidents as they continue to improve; why would one such incident, that wasn't even fatal, have such a dramatic response?
- Sure it's possible that Cruise was just a row in a spreadsheet to some finance higher-ups at GM, and it was dropped like that because it wasn't making them money yet. But to drop such a multi-year investment has to have been something else, a deeper problem:
- E.g. Did they realize/decide that the technical architecture was somehow flawed and there was an unbreakable barrier that prevents them from achieving that remaining 10%?"},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: What really happened to Cruise self-driving cars?"}},"_tags":["story","author_eh_why_not","story_43888265","ask_hn"],"author":"eh_why_not","children":[43888336,43888602,43896015],"created_at":"2025-05-04T17:59:23Z","created_at_i":1746381563,"num_comments":4,"objectID":"43888265","points":4,"story_id":43888265,"story_text":"Now that plenty of time has passed, I'm wondering if someone here had an inside track and is willing to share their insights.
Rough summary of the available public information:
- They had an accident in 2023 injuring, by not stopping instantly, a passerby who was hit by another car.
- They were investigated and then California temporarily suspended their license to drive around.
- By the end of 2023 there was a massive layoff, including top executives leaving. There were claims that the operation was too human-dependent / required too much human interventions.
- In 2024 there were some news that they were starting up again (outside CA maybe).
- In Dec 2024, GM dropped their funding altogether and gave up on self-driving / robotaxi, refocusing on driver assistance instead. Some analysts claimed that was in part (in addition to finances) due to politics and them not wanting to face off with Tesla under the new administration.
It always seemed to me that something was missing from the overall picture:
- While second to Waymo, they still seemed already far ahead from others; maybe they were, say, 90% there.
- Others, like Tesla, had many accidents, including fatal ones. But for those the gauntlet was not thrown that quickly. It is expected that self-driving cars will have accidents as they continue to improve; why would one such incident, that wasn't even fatal, have such a dramatic response?
- Sure it's possible that Cruise was just a row in a spreadsheet to some finance higher-ups at GM, and it was dropped like that because it wasn't making them money yet. But to drop such a multi-year investment has to have been something else, a deeper problem:
- E.g. Did they realize/decide that the technical architecture was somehow flawed and there was an unbreakable barrier that prevents them from achieving that remaining 10%?","title":"Ask HN: What really happened to Cruise self-driving cars?","updated_at":"2025-05-06T04:39:03Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"petesoder"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"I couldn't agree more. Engineers are loathe to enter into a traditional recruiting process (i.e. go in through the front door) when they have so many options out there.
It's far more attractive if they can 'window shop' and get a true sense for your team/culture/challenges without kicking off a gauntlet of interviews for them to learn the same stuff.
Obviously, interviews are still important for the hiring team, but it's time to make the overall process better (and more front-loaded with the company's engineering project info) for the engineer candidates."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Why PR for Your Engineering Team is not Crazy Talk"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"http://blog.petesoder.com/blog/2013/09/23/why-engineering-pr-is-not-crazy-talk/"}},"_tags":["comment","author_petesoder","story_6440158"],"author":"petesoder","comment_text":"I couldn't agree more. Engineers are loathe to enter into a traditional recruiting process (i.e. go in through the front door) when they have so many options out there.
It's far more attractive if they can 'window shop' and get a true sense for your team/culture/challenges without kicking off a gauntlet of interviews for them to learn the same stuff.
Obviously, interviews are still important for the hiring team, but it's time to make the overall process better (and more front-loaded with the company's engineering project info) for the engineer candidates.","created_at":"2013-09-24T21:08:00Z","created_at_i":1380056880,"objectID":"6440770","parent_id":6440682,"story_id":6440158,"story_title":"Why PR for Your Engineering Team is not Crazy Talk","story_url":"http://blog.petesoder.com/blog/2013/09/23/why-engineering-pr-is-not-crazy-talk/","updated_at":"2024-09-19T20:00:27Z"},{"_highlightResult":{"author":{"fullyHighlighted":true,"matchLevel":"partial","matchedWords":["ghuntley"],"value":"ghuntley"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"Just wrote up a quickie about how Nix + LLMs make patching everything super simple via overlays. Simplest security response I've ever done in my life. http://ghuntley.com/patch"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"EMudhra Intermediate CA Compromise"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://emudhra.we-have-a-problem.com/"}},"_tags":["comment","author_ghuntley","story_50166579"],"author":"ghuntley","comment_text":"Just wrote up a quickie about how Nix + LLMs make patching everything super simple via overlays. Simplest security response I've ever done in my life. http://ghuntley.com/patch","created_at":"2026-10-20T16:44:32Z","created_at_i":1792514672,"objectID":"50169111","parent_id":50166579,"story_id":50166579,"story_title":"EMudhra Intermediate CA Compromise","story_url":"https://emudhra.we-have-a-problem.com/","updated_at":"2026-10-20T21:27:32Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"paul7986"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"As a consumer paying $20 now to chatGPT where i hit those gauntlet usage messages then using Muse a lot and not receiving any such messages. Im questioning my $20 a month GPT subscription and my $5 a month Gemini (Google has enough money as Meta to make it free) subscription.
Overall Ai is doing all the software development now and we the consumers are paying for this while many are out of work and if you are out of work the US tech job market is brutal!
Ai needs to be free to use without limits and in time Im guessing it will be. That could be Meta and or another who accomplishes this. Whoever does wins this race."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"ChatGPT Pro 500"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://help.openai.com/en/articles/9793128-about-chatgpt-pro-tiers"}},"_tags":["comment","author_paul7986","story_49896975"],"author":"paul7986","children":[49898594],"comment_text":"As a consumer paying $20 now to chatGPT where i hit those gauntlet usage messages then using Muse a lot and not receiving any such messages. Im questioning my $20 a month GPT subscription and my $5 a month Gemini (Google has enough money as Meta to make it free) subscription.
Overall Ai is doing all the software development now and we the consumers are paying for this while many are out of work and if you are out of work the US tech job market is brutal!
Ai needs to be free to use without limits and in time Im guessing it will be. That could be Meta and or another who accomplishes this. Whoever does wins this race.","created_at":"2026-09-29T18:22:09Z","created_at_i":1790706129,"objectID":"49898038","parent_id":49897185,"story_id":49896975,"story_title":"ChatGPT Pro 500","story_url":"https://help.openai.com/en/articles/9793128-about-chatgpt-pro-tiers","updated_at":"2026-09-29T19:31:10Z"},{"_highlightResult":{"author":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ghuntley"],"value":"gruntled-worker"},"comment_text":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["overalls"],"value":"Another strategy that's often used is to cause motion then detect the resulting transmission spike. That's also defeatable of course (cache+burst, CBR, etc).
The ideal detector should incorporate multiple methods. Someone should make a site that rates these for each detection method then adds them into an overall score."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Smartphone LED detects hidden cameras with AI"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.chosun.com/english/industry-en/2026/08/30/SBFXUIJQYZEARKP5T4FBAY25HQ/"}},"_tags":["comment","author_gruntled-worker","story_49496292"],"author":"gruntled-worker","comment_text":"Another strategy that's often used is to cause motion then detect the resulting transmission spike. That's also defeatable of course (cache+burst, CBR, etc).
The ideal detector should incorporate multiple methods. Someone should make a site that rates these for each detection method then adds them into an overall score.","created_at":"2026-09-01T03:25:40Z","created_at_i":1788233140,"objectID":"49517632","parent_id":49496292,"story_id":49496292,"story_title":"Smartphone LED detects hidden cameras with AI","story_url":"https://www.chosun.com/english/industry-en/2026/08/30/SBFXUIJQYZEARKP5T4FBAY25HQ/","updated_at":"2026-09-01T18:09:57Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"strken"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"I took a Greyhound a couple of times when I was in the US, and the experience immediately showed me why Americans hate buses and coaches.
My first transfer was in Sacramento. The entire bus got held up for over an hour because someone saw a man with a knife and security had to search absolutely everyone to try to find it.
Half the stations were literally crumbling, as in the ceilings were falling down and covered in water stains and flecks of black mould. The drivers often turned up hours late, which is apparently expected and normal. The stations tended to be in exciting hotspots such as Skid Row, to cater for the desperate clientele who had no choice but to run the gauntlet.
Also, after the first time I rode it and told everyone about the knife that nobody ever found, people started showing me news stories about the man who got beheaded on a Greyhound in Canada.
Overall I think they have very patchy bus and coach systems and over-index on the worst examples."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Bus stop balancing is fast, cheap, and effective"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://worksinprogress.co/issue/the-united-states-needs-fewer-bus-stops/"}},"_tags":["comment","author_strken","story_47153798"],"author":"strken","children":[47158682],"comment_text":"I took a Greyhound a couple of times when I was in the US, and the experience immediately showed me why Americans hate buses and coaches.
My first transfer was in Sacramento. The entire bus got held up for over an hour because someone saw a man with a knife and security had to search absolutely everyone to try to find it.
Half the stations were literally crumbling, as in the ceilings were falling down and covered in water stains and flecks of black mould. The drivers often turned up hours late, which is apparently expected and normal. The stations tended to be in exciting hotspots such as Skid Row, to cater for the desperate clientele who had no choice but to run the gauntlet.
Also, after the first time I rode it and told everyone about the knife that nobody ever found, people started showing me news stories about the man who got beheaded on a Greyhound in Canada.
Overall I think they have very patchy bus and coach systems and over-index on the worst examples.","created_at":"2026-02-25T21:32:19Z","created_at_i":1772055139,"objectID":"47158184","parent_id":47157042,"story_id":47153798,"story_title":"Bus stop balancing is fast, cheap, and effective","story_url":"https://worksinprogress.co/issue/the-united-states-needs-fewer-bus-stops/","updated_at":"2026-03-05T23:37:25Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"pwg"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"Part of the point of the split when the US Congress was designed was to intentionally make it difficult for bills to pass, because they had to pass votes in two independent houses, that (presumably) were focused on differing agendas.
This inherent difficulty was the intended outcome to try to assure that only bills which had strong support overall from different perspectives and viewpoints would make it through the double gauntlet."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Hemp ban hidden inside government shutdown bill"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://hightimes.com/news/politics/hemp-ban-hidden-inside-government-shutdown-bill/"}},"_tags":["comment","author_pwg","story_45916152"],"author":"pwg","children":[45920385],"comment_text":"Part of the point of the split when the US Congress was designed was to intentionally make it difficult for bills to pass, because they had to pass votes in two independent houses, that (presumably) were focused on differing agendas.
This inherent difficulty was the intended outcome to try to assure that only bills which had strong support overall from different perspectives and viewpoints would make it through the double gauntlet.","created_at":"2025-11-13T20:36:33Z","created_at_i":1763066193,"objectID":"45920185","parent_id":45917750,"story_id":45916152,"story_title":"Hemp ban hidden inside government shutdown bill","story_url":"https://hightimes.com/news/politics/hemp-ban-hidden-inside-government-shutdown-bill/","updated_at":"2026-03-05T22:58:49Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"steveplace"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"I went to a very large public university. The engineering program was so-so, and they valued ABET over actual learning.
Senior Design was known as the \"gauntlet,\" not necessarily due to the actual building of something, but for the 100-page research paper that they required. They assumed that if each member of the team writes about 25 pages, then they must understand the underlying concepts.
This didn't affect me as much as it did my team. As you can tell, my long-windedness knows no bounds when behind a keyboard. My team however had the collective grammatical know-how of [insert punchline here].
It was my job to proofread the paper. It was straight copy and paste. Utter bullshit.
And I understand that it was my ethical obligation to correct this, if not report it. But it was clearly stated at the beginning of the class: \"if one fails, you all fail.\" I was not prepared to deal with that. So I compromised my standards and just finished the assignment.
I knew that there wouldn't be repercussions; all they do is count the number of pages, make sure there isn't too much whitespace, and that there's no more than 50% of space devoted to pictures.
What can be done? With this particular example there was a large discrepancy with the overall goals that should have been, the goals that were, and the metrics that they measured it."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Has anybody observed cheating?"}},"_tags":["comment","author_steveplace","story_455637"],"author":"steveplace","comment_text":"I went to a very large public university. The engineering program was so-so, and they valued ABET over actual learning.
Senior Design was known as the \"gauntlet,\" not necessarily due to the actual building of something, but for the 100-page research paper that they required. They assumed that if each member of the team writes about 25 pages, then they must understand the underlying concepts.
This didn't affect me as much as it did my team. As you can tell, my long-windedness knows no bounds when behind a keyboard. My team however had the collective grammatical know-how of [insert punchline here].
It was my job to proofread the paper. It was straight copy and paste. Utter bullshit.
And I understand that it was my ethical obligation to correct this, if not report it. But it was clearly stated at the beginning of the class: \"if one fails, you all fail.\" I was not prepared to deal with that. So I compromised my standards and just finished the assignment.
I knew that there wouldn't be repercussions; all they do is count the number of pages, make sure there isn't too much whitespace, and that there's no more than 50% of space devoted to pictures.
What can be done? With this particular example there was a large discrepancy with the overall goals that should have been, the goals that were, and the metrics that they measured it.","created_at":"2009-01-29T02:59:15Z","created_at_i":1233197955,"objectID":"455696","parent_id":455637,"story_id":455637,"story_title":"Ask HN: Has anybody observed cheating?","updated_at":"2024-09-19T16:32:06Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"a022311"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"Awesome work! I've been dreaming of this project myself with a very similar stack (basically React instead of Svelte), but I'm very excited to finally have a Raycast alternative for Linux that has both consistent UI and extension compatibility.
The most impressive part is probably your age, because this isn't an easy project even for senior devs!
I haven't tried it yet, but I can't wait to find some time for that.
I've researched applications like this for Linux quite extensively and I think you might find the following tips of interest:\n- For the slow extension startup issue you mentioned, consider Deno as a runtime as it has a better sandbox and is faster than Node overall. There may be some compatibility issues, but if I remember correctly most stuff is handled by the special Raycast extension libraries which you implement manually anyway.\n- I'd consider Numbat [0] for replacing the calculator implementation you have now. As far as I can tell, it should have feature parity with SoulverCore and it's also written in Rust, so interfacing with it should be much easier and won't require the FFI work you're doing now.\n- Project Gauntlet [1] is another project which has gotten quite close to implementing a full-featured Raycast alternative and might be worth taking inspiration from. It would certainly be very cool if you can make the UI rendering native at some point (although I guess Rust isn't perfect for native UI at the moment [2])
Keep up the good work!
[0]: https://github.com/sharkdp/numbat
[1]: https://github.com/project-gauntlet/gauntlet
[2]: https://areweguiyet.com/"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: A Raycast-compatible launcher for Linux"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/ByteAtATime/raycast-linux"}},"_tags":["comment","author_a022311","story_44551762"],"author":"a022311","comment_text":"Awesome work! I've been dreaming of this project myself with a very similar stack (basically React instead of Svelte), but I'm very excited to finally have a Raycast alternative for Linux that has both consistent UI and extension compatibility.
The most impressive part is probably your age, because this isn't an easy project even for senior devs!
I haven't tried it yet, but I can't wait to find some time for that.
I've researched applications like this for Linux quite extensively and I think you might find the following tips of interest:\n- For the slow extension startup issue you mentioned, consider Deno as a runtime as it has a better sandbox and is faster than Node overall. There may be some compatibility issues, but if I remember correctly most stuff is handled by the special Raycast extension libraries which you implement manually anyway.\n- I'd consider Numbat [0] for replacing the calculator implementation you have now. As far as I can tell, it should have feature parity with SoulverCore and it's also written in Rust, so interfacing with it should be much easier and won't require the FFI work you're doing now.\n- Project Gauntlet [1] is another project which has gotten quite close to implementing a full-featured Raycast alternative and might be worth taking inspiration from. It would certainly be very cool if you can make the UI rendering native at some point (although I guess Rust isn't perfect for native UI at the moment [2])
Keep up the good work!
[0]: https://github.com/sharkdp/numbat
[1]: https://github.com/project-gauntlet/gauntlet
[2]: https://areweguiyet.com/","created_at":"2025-07-13T19:12:06Z","created_at_i":1752433926,"objectID":"44552699","parent_id":44551762,"story_id":44551762,"story_title":"Show HN: A Raycast-compatible launcher for Linux","story_url":"https://github.com/ByteAtATime/raycast-linux","updated_at":"2025-07-29T15:15:51Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"skymt"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"I can think of a couple arcade examples analogous to the \"in-game currency\" variety of IAP. In Gauntlet, inserting coins adds to the player's health, which constantly ticks down. In Lunar Lander, inserting coins buys fuel, which is needed to maneuver.
By the time arcade games were sophisticated enough to have things like multiple weapons and super moves, developers had settled on the model where extra coins are used to buy the ability to continue a game after failing. The arcade industry is quite conservative overall."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"The fall of Angry Birds"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"http://www.treysmithblog.com/the-fall-of-angry-birds/"}},"_tags":["comment","author_skymt","story_4432086"],"author":"skymt","comment_text":"I can think of a couple arcade examples analogous to the \"in-game currency\" variety of IAP. In Gauntlet, inserting coins adds to the player's health, which constantly ticks down. In Lunar Lander, inserting coins buys fuel, which is needed to maneuver.
By the time arcade games were sophisticated enough to have things like multiple weapons and super moves, developers had settled on the model where extra coins are used to buy the ability to continue a game after failing. The arcade industry is quite conservative overall.","created_at":"2012-08-25T16:56:52Z","created_at_i":1345913812,"objectID":"4432533","parent_id":4432419,"story_id":4432086,"story_title":"The fall of Angry Birds","story_url":"http://www.treysmithblog.com/the-fall-of-angry-birds/","updated_at":"2024-09-19T18:47:09Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"bbor"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"A) Looks really good, will be checking it out in depth as I get time! Thanks for sharing.
B) The endorsements are interesting before you even get to the book; I know all textbooks are marketed, but this seems like quite the concerted effort. For example, take Judea Pearl's quote (an under-appreciated giant):
This revised and extended edition of Artificial Intelligence: Foundations of Computational Agents should become the standard text of AI education.\n(Quick context for those rusty on academic AI: Russell & Norvig's 1995 (4th ed in 2020) AI: A Modern Approach ("AIAMA") is the de facto book for AI survey courses, supposedly used in 1500 universities via 9 languages as of 2023.[1])
I might be reading drama into the situation that isn't necessary, but it sure looks like they're trying to establish a connectionist/"scruffy", ML-based, Python-first replacement for AIAMA's symbolic/"neat", logic-based, Lisp-first approach. The 1st ed hit desks in 2010, and the endorsements are overwhelmingly from scruffy scientists & engineers. Obviously, this mirrors the industry's overall trend[2]... at this point, most laypeople think AI is ML. Nice to see a more nuanced--yet still scruffy-forward--approach gaining momentum; even Gary Marcus is on board, a noted Neat!
C) ...Ok, after writing an already-long comment (sorry) I did a quantitative comparison of the two books, which I figured y'all might find interesting! I'll link a screenshot[3] and the Google Sheet itself[4] below, but here's some highlights b/w "AMA" (the reigning champion) and "FCA" (the scrappy challenger):
1. My thesis was definitely correct; by my subjective estimation, AMA is ~6:3 neat:scruffy (57%:32%), vs. a ~3:5 ratio for FCA (34%:50%).
2. My second thesis is also seemingly correct: FCA dedicates the last few pages of every section to "Social Impact", aka ethics. Both books discuss the topic in more depth in the conclusion, representing ~4% of each.
3. FCA seems to have some significant pedagogical advantages, namely length (797 pages vs. AMA's 1023 pages) and the inclusion of in-text exercises at the end of every section.
4. Both publish source code in multiple languages, but AMA had to be ported to Python from Lisp, whereas FCA is natively in Python (which, obviously, dominates AI atm). The FCA authors actually wrote their own "psuedo-code" Python library, which is both concerning and potentially helpful.
5. Finally, FCA includes sections explicitly focused on data structures, rather than just weaving them into discussions of algorithms & behavioral patterns. I for one think this is a really great idea, and where I see most short-term advances in unified (symbolic + stochastic) AI research coming from! Lots of gold to be mined in 75 years of thought.
Apologies, as always, for the long comment -- my only solace is that you can quickly minimize it. I should start a blog where I can muse to my heart's content...
TL;DR: This new book is shorter, more ML-centric, and arguably uses more modern pedagogical techniques; in general, it seems to be a slightly more engineer-focused answer to Russell & Norvig's more academic-focused standard text.
[1] AIAMA: https://en.wikipedia.org/wiki/Artificial_Intelligence:_A_Mod...
[2] NGRAM: https://books.google.com/ngrams/graph?content=%28Machine+Lea...
[3] Screenshot: https://imgur.com/a/x8QMbno
[4] Google Sheet: https://docs.google.com/spreadsheets/d/1Gw9lxWhhTxjjTstyAKli..."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Artificial Intelligence: Foundations of Computational Agents"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://artint.info/index.html"}},"_tags":["comment","author_bbor","story_43378226"],"author":"bbor","children":[43409270,43409373],"comment_text":"A) Looks really good, will be checking it out in depth as I get time! Thanks for sharing.
B) The endorsements are interesting before you even get to the book; I know all textbooks are marketed, but this seems like quite the concerted effort. For example, take Judea Pearl's quote (an under-appreciated giant):
This revised and extended edition of Artificial Intelligence: Foundations of Computational Agents should become the standard text of AI education.\n(Quick context for those rusty on academic AI: Russell & Norvig's 1995 (4th ed in 2020) AI: A Modern Approach ("AIAMA") is the de facto book for AI survey courses, supposedly used in 1500 universities via 9 languages as of 2023.[1])
I might be reading drama into the situation that isn't necessary, but it sure looks like they're trying to establish a connectionist/"scruffy", ML-based, Python-first replacement for AIAMA's symbolic/"neat", logic-based, Lisp-first approach. The 1st ed hit desks in 2010, and the endorsements are overwhelmingly from scruffy scientists & engineers. Obviously, this mirrors the industry's overall trend[2]... at this point, most laypeople think AI is ML. Nice to see a more nuanced--yet still scruffy-forward--approach gaining momentum; even Gary Marcus is on board, a noted Neat!
C) ...Ok, after writing an already-long comment (sorry) I did a quantitative comparison of the two books, which I figured y'all might find interesting! I'll link a screenshot[3] and the Google Sheet itself[4] below, but here's some highlights b/w "AMA" (the reigning champion) and "FCA" (the scrappy challenger):
1. My thesis was definitely correct; by my subjective estimation, AMA is ~6:3 neat:scruffy (57%:32%), vs. a ~3:5 ratio for FCA (34%:50%).
2. My second thesis is also seemingly correct: FCA dedicates the last few pages of every section to "Social Impact", aka ethics. Both books discuss the topic in more depth in the conclusion, representing ~4% of each.
3. FCA seems to have some significant pedagogical advantages, namely length (797 pages vs. AMA's 1023 pages) and the inclusion of in-text exercises at the end of every section.
4. Both publish source code in multiple languages, but AMA had to be ported to Python from Lisp, whereas FCA is natively in Python (which, obviously, dominates AI atm). The FCA authors actually wrote their own "psuedo-code" Python library, which is both concerning and potentially helpful.
5. Finally, FCA includes sections explicitly focused on data structures, rather than just weaving them into discussions of algorithms & behavioral patterns. I for one think this is a really great idea, and where I see most short-term advances in unified (symbolic + stochastic) AI research coming from! Lots of gold to be mined in 75 years of thought.
Apologies, as always, for the long comment -- my only solace is that you can quickly minimize it. I should start a blog where I can muse to my heart's content...
TL;DR: This new book is shorter, more ML-centric, and arguably uses more modern pedagogical techniques; in general, it seems to be a slightly more engineer-focused answer to Russell & Norvig's more academic-focused standard text.
[1] AIAMA: https://en.wikipedia.org/wiki/Artificial_Intelligence:_A_Mod...
[2] NGRAM: https://books.google.com/ngrams/graph?content=%28Machine+Lea...
[3] Screenshot: https://imgur.com/a/x8QMbno
[4] Google Sheet: https://docs.google.com/spreadsheets/d/1Gw9lxWhhTxjjTstyAKli...","created_at":"2025-03-19T05:37:53Z","created_at_i":1742362673,"objectID":"43408557","parent_id":43378226,"story_id":43378226,"story_title":"Artificial Intelligence: Foundations of Computational Agents","story_url":"https://artint.info/index.html","updated_at":"2025-03-22T21:17:47Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"OldGuyInTheClub"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"Pretty bad, overall. I got diddlysquat out of LinkedIn when I was job hunting 10+ years ago but then I got nothing out of my other leads, either. I was old back then and older now so I am staying at my employer until I'm fired, retired, or expired. LI may be required these days to get anywhere, I just adjusted my travel plans so I didn't need to deal with its BS or the general Talent Acq/HR gauntlet required to get past square one. Probably not helpful to you but there it is..."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Life lessons from the first half-century of my career"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://cacm.acm.org/opinion/life-lessons-from-the-first-half-century-of-my-career/"}},"_tags":["comment","author_OldGuyInTheClub","story_42795646"],"author":"OldGuyInTheClub","comment_text":"Pretty bad, overall. I got diddlysquat out of LinkedIn when I was job hunting 10+ years ago but then I got nothing out of my other leads, either. I was old back then and older now so I am staying at my employer until I'm fired, retired, or expired. LI may be required these days to get anywhere, I just adjusted my travel plans so I didn't need to deal with its BS or the general Talent Acq/HR gauntlet required to get past square one. Probably not helpful to you but there it is...","created_at":"2025-01-25T19:10:52Z","created_at_i":1737832252,"objectID":"42823793","parent_id":42823718,"story_id":42795646,"story_title":"Life lessons from the first half-century of my career","story_url":"https://cacm.acm.org/opinion/life-lessons-from-the-first-half-century-of-my-career/","updated_at":"2025-01-25T19:16:23Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"tunesmith"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"This is fun, and holds true in the creative writing group I run. We use a website I programmed that helps us collaborate on writing branching fiction. We have a mapping utility that creates graphs like in the article, except more animated (d3.js, elkjs).
As different authors can start their own new stories, one thing I often have to deal with is that they want to design their story to have both long path lengths (multiple chapters before an ending), and also high choice count. Those of you who know something about geometric series know that this causes problems. I often have to tell them they can't have everything they want, which causes minor drama. :)
As a result, one of our stories basically shot its "choice budget" in the first few chapters, leading to many linear paths in the latter parts of the narratives, which is fun in its own way.
Another of our stories has just started playing with the "gauntlet pattern" as the article describes. For this one, we decided that all chapters must be in the "same universe", just following different characters' perspectives, and are planning for certain "anchor chapters" where all characters come together for a meeting. Probably the detective questioning them as a group (it's a murder mystery).
All of our stories are supposed to be literary, so usually in third person, sometimes first, never the second-person. So we don't tend to use choices and chapters as directions and rooms; it's all about how the plot moves. We also don't track state; they're designed to be able to be printed as books people can page through.
Overall a super-fun project for me and a handful of other writers, it's been a consistent way to spend a few hours of fun each week."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Standard patterns in choice-based games (2015)"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://heterogenoustasks.wordpress.com/2015/01/26/standard-patterns-in-choice-based-games/"}},"_tags":["comment","author_tunesmith","story_42678647"],"author":"tunesmith","children":[42681051],"comment_text":"This is fun, and holds true in the creative writing group I run. We use a website I programmed that helps us collaborate on writing branching fiction. We have a mapping utility that creates graphs like in the article, except more animated (d3.js, elkjs).
As different authors can start their own new stories, one thing I often have to deal with is that they want to design their story to have both long path lengths (multiple chapters before an ending), and also high choice count. Those of you who know something about geometric series know that this causes problems. I often have to tell them they can't have everything they want, which causes minor drama. :)
As a result, one of our stories basically shot its "choice budget" in the first few chapters, leading to many linear paths in the latter parts of the narratives, which is fun in its own way.
Another of our stories has just started playing with the "gauntlet pattern" as the article describes. For this one, we decided that all chapters must be in the "same universe", just following different characters' perspectives, and are planning for certain "anchor chapters" where all characters come together for a meeting. Probably the detective questioning them as a group (it's a murder mystery).
All of our stories are supposed to be literary, so usually in third person, sometimes first, never the second-person. So we don't tend to use choices and chapters as directions and rooms; it's all about how the plot moves. We also don't track state; they're designed to be able to be printed as books people can page through.
Overall a super-fun project for me and a handful of other writers, it's been a consistent way to spend a few hours of fun each week.","created_at":"2025-01-13T04:45:41Z","created_at_i":1736743541,"objectID":"42680160","parent_id":42678647,"story_id":42678647,"story_title":"Standard patterns in choice-based games (2015)","story_url":"https://heterogenoustasks.wordpress.com/2015/01/26/standard-patterns-in-choice-based-games/","updated_at":"2025-01-25T23:00:53Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"tivert"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"> Based on my experience with doctors medicine is not one of those fields. Residency is the final gauntlet that matters and that\u2019s the part that\u2019s artificially restricted.
You're getting the basic facts wrong. As least in the US, the restriction is acceptance to medical school. Part of the reasoning is they don't want to accept anyone who they think will fail years down the line, because that's bad for everyone (you, me, the failed student, and the teachers who aren't in it for the $$$). There's a surplus of residency slots (overall), because otherwise no foreign-trained doctors would ever be able to practice here.
Lots of foreign-trained doctors practice here.
My position is they need to increase medical school admissions (and residency slots), but not throw the doors wide open to all comers. The slots need to be managed such that anyone who gets can complete the program and will have a job waiting for them that pays commensurate with the amount of training they receive, but probably much less than many doctors get now.
> I mean really, where do we stop? There\u2019s many more English literature graduates than jobs. Same with ____ studies and just about every other humanities degree.
Yes. What's happening with English literature graduates (at least on the PhD level) is an exploitative travesty.
> ...but restricting education based on job supply is just harmful to societies on so many levels.
I disagree. It's harmful to string people along with false hope, only to slam the door in their face at the end. It's harmful for society for people to waste time on a job training program (which is what medical education really is) for a job they'll never get.
There needs to be a distinction made between general education and job training. Unless you're independently wealthy, specialized formal education needs to perform both functions.
> The washout rate for A&P aircraft mechanics is on the order of 75%.
Aircraft mechanic school is about 18-24 months (https://www.indeed.com/career-advice/finding-a-job/how-long-...), and costs a few tens of thousands of dollars. That's not comparable, at all.
> Imagine if healthcare were as reliable as airplanes (though there\u2019s a shortage there too).
Sure imagine that, but people aren't manufactured products. Not a valid comparison."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Florida Eases Licensing Requirements for Foreign Trained Doctors"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.murthy.com/2024/10/17/florida-eases-licensing-requirements-for-foreign-trained-doctors/"}},"_tags":["comment","author_tivert","story_41942023"],"author":"tivert","comment_text":"> Based on my experience with doctors medicine is not one of those fields. Residency is the final gauntlet that matters and that\u2019s the part that\u2019s artificially restricted.
You're getting the basic facts wrong. As least in the US, the restriction is acceptance to medical school. Part of the reasoning is they don't want to accept anyone who they think will fail years down the line, because that's bad for everyone (you, me, the failed student, and the teachers who aren't in it for the $$$). There's a surplus of residency slots (overall), because otherwise no foreign-trained doctors would ever be able to practice here.
Lots of foreign-trained doctors practice here.
My position is they need to increase medical school admissions (and residency slots), but not throw the doors wide open to all comers. The slots need to be managed such that anyone who gets can complete the program and will have a job waiting for them that pays commensurate with the amount of training they receive, but probably much less than many doctors get now.
> I mean really, where do we stop? There\u2019s many more English literature graduates than jobs. Same with ____ studies and just about every other humanities degree.
Yes. What's happening with English literature graduates (at least on the PhD level) is an exploitative travesty.
> ...but restricting education based on job supply is just harmful to societies on so many levels.
I disagree. It's harmful to string people along with false hope, only to slam the door in their face at the end. It's harmful for society for people to waste time on a job training program (which is what medical education really is) for a job they'll never get.
There needs to be a distinction made between general education and job training. Unless you're independently wealthy, specialized formal education needs to perform both functions.
> The washout rate for A&P aircraft mechanics is on the order of 75%.
Aircraft mechanic school is about 18-24 months (https://www.indeed.com/career-advice/finding-a-job/how-long-...), and costs a few tens of thousands of dollars. That's not comparable, at all.
> Imagine if healthcare were as reliable as airplanes (though there\u2019s a shortage there too).
Sure imagine that, but people aren't manufactured products. Not a valid comparison.","created_at":"2024-10-25T16:58:31Z","created_at_i":1729875511,"objectID":"41947148","parent_id":41942733,"story_id":41942023,"story_title":"Florida Eases Licensing Requirements for Foreign Trained Doctors","story_url":"https://www.murthy.com/2024/10/17/florida-eases-licensing-requirements-for-foreign-trained-doctors/","updated_at":"2024-10-25T17:31:00Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"decafninja"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"It\u2019s always been recruiters reaching out to me. I can\u2019t see it being much different now.
I\u2019ve never, ever, gotten a human response from any company that I cold applied. Either ignored, or get the automated rejection email.
Recruiters that I could cold contact usually tell me thanks for reaching out, and that they\u2019ll be in touch if they\u2019re interested. They never are.
I could probably get -a- job through networking, but probably not a job I\u2019d be very interested in. I have a network, but very few people work somewhere that I\u2019d like to work at. The few that do would at best guarantee I get fed into the leetcode pipeline and then it would all be on me to leetcode my way in.
I know some master networkers that can get offers and switch jobs via their network at the drop of a hat. Again, the catch here is that they have to be very unpicky. Overall, many of the best places to work are gatekeeped with leetcode.
I\u2019m currently at what I consider an \u201cendgame\u201d company. I wouldn\u2019t mind if this is the last company I work at (I\u2019m 40). If I can last here until I\u2019m late 40s or early 50s, then I\u2019d rather semi-retire instead of putting myself through the leetcode gauntlet again."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: How do you find employment opportunities in 2024?"}},"_tags":["comment","author_decafninja","story_39565788"],"author":"decafninja","children":[39570179,39572577],"comment_text":"It\u2019s always been recruiters reaching out to me. I can\u2019t see it being much different now.
I\u2019ve never, ever, gotten a human response from any company that I cold applied. Either ignored, or get the automated rejection email.
Recruiters that I could cold contact usually tell me thanks for reaching out, and that they\u2019ll be in touch if they\u2019re interested. They never are.
I could probably get -a- job through networking, but probably not a job I\u2019d be very interested in. I have a network, but very few people work somewhere that I\u2019d like to work at. The few that do would at best guarantee I get fed into the leetcode pipeline and then it would all be on me to leetcode my way in.
I know some master networkers that can get offers and switch jobs via their network at the drop of a hat. Again, the catch here is that they have to be very unpicky. Overall, many of the best places to work are gatekeeped with leetcode.
I\u2019m currently at what I consider an \u201cendgame\u201d company. I wouldn\u2019t mind if this is the last company I work at (I\u2019m 40). If I can last here until I\u2019m late 40s or early 50s, then I\u2019d rather semi-retire instead of putting myself through the leetcode gauntlet again.","created_at":"2024-03-02T03:37:58Z","created_at_i":1709350678,"objectID":"39569577","parent_id":39565788,"story_id":39565788,"story_title":"Ask HN: How do you find employment opportunities in 2024?","updated_at":"2024-09-20T16:32:04Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"roenxi"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"I'd estimate that at least 1/1000 people are capable of running this gauntlet; so in my opinion this is a little pessimistic. That being said, the gentleman is hitting the message like a carpenter in a hurry.
> ... the vast majority of people being leeches ...
Yeah, but unfairly pejorative. Babies and old people are technically leeches too but we love them and shouldn't call them that.
It is fairer to be accurate and say most people don't create much in the way of resources. You can see this play out pretty straightforwardly if you assume a dollar in savings represents about a dollar of value created that wasn't then consumed. Most people struggle to get that metric positive.
That is fine though. Capitalism forces people not to be outrageous resource drains and even tilts the scales in the micro so that someone with no savings still probably made a positive contribution to the overall wealth of the nation."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Several Truths Regarding Success"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://risingthumb.xyz/Writing/Blog/TruthsAboutSuccess"}},"_tags":["comment","author_roenxi","story_39174577"],"author":"roenxi","children":[39175116,39175154,39175166,39175186,39175228],"comment_text":"I'd estimate that at least 1/1000 people are capable of running this gauntlet; so in my opinion this is a little pessimistic. That being said, the gentleman is hitting the message like a carpenter in a hurry.
> ... the vast majority of people being leeches ...
Yeah, but unfairly pejorative. Babies and old people are technically leeches too but we love them and shouldn't call them that.
It is fairer to be accurate and say most people don't create much in the way of resources. You can see this play out pretty straightforwardly if you assume a dollar in savings represents about a dollar of value created that wasn't then consumed. Most people struggle to get that metric positive.
That is fine though. Capitalism forces people not to be outrageous resource drains and even tilts the scales in the micro so that someone with no savings still probably made a positive contribution to the overall wealth of the nation.","created_at":"2024-01-29T11:13:48Z","created_at_i":1706526828,"objectID":"39175047","parent_id":39174577,"story_id":39174577,"story_title":"Several Truths Regarding Success","story_url":"https://risingthumb.xyz/Writing/Blog/TruthsAboutSuccess","updated_at":"2025-08-14T22:46:38Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"digital-cygnet"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"What I don't understand is the bit at the end (spoilers): he finds out that a person who Kraken's fraud team sees as a victim of scams, but hasn't been able to contact, has called his Gauntlet. He's then able to reach out and help her, as well as other scam victims who end up in the Gauntlet.
So, three questions -
1. How do scam victims end up in the Gauntlet at all? I thought the idea was that Kitboga and his team pose as marks, send the bogus QR code to the scammer, and that's their whole pipeline. How do legitimate people end up in there?
2. Assuming the above is something like "scammers are clearly manipulating scam victims into helping them with the Gauntlet", doesn't that raise questions about the glee with which he, and all of us, are watching "scammers'" frustration? It becomes a more nuanced moral calculus if some number of the people you are frustrating are innocent people manipulated into navigating this system for scammers. You could argue it's still net good because otherwise the effort spent manipulating them would have been spent doing a real scam on them, but honestly I'm not sure
3. How did they identify the non-scammers who ended up on the platform? If there was a solid answer to this I suppose it could mitigate (2), but it's hard for me to believe (unless it's something very labor intensive that would make the automated nature of the system less useful)
Overall I still enjoyed the video and found parts completely hilarious (and far too realistic given my own experience on phone trees). But the above does give me pause about unreserved support for what he's doing"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"I trapped scammers in an impossible maze [video]"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.youtube.com/watch?v=dWzz3NeDz3E"}},"_tags":["comment","author_digital-cygnet","story_38130418"],"author":"digital-cygnet","children":[38136676,38137894,38162247],"comment_text":"What I don't understand is the bit at the end (spoilers): he finds out that a person who Kraken's fraud team sees as a victim of scams, but hasn't been able to contact, has called his Gauntlet. He's then able to reach out and help her, as well as other scam victims who end up in the Gauntlet.
So, three questions -
1. How do scam victims end up in the Gauntlet at all? I thought the idea was that Kitboga and his team pose as marks, send the bogus QR code to the scammer, and that's their whole pipeline. How do legitimate people end up in there?
2. Assuming the above is something like "scammers are clearly manipulating scam victims into helping them with the Gauntlet", doesn't that raise questions about the glee with which he, and all of us, are watching "scammers'" frustration? It becomes a more nuanced moral calculus if some number of the people you are frustrating are innocent people manipulated into navigating this system for scammers. You could argue it's still net good because otherwise the effort spent manipulating them would have been spent doing a real scam on them, but honestly I'm not sure
3. How did they identify the non-scammers who ended up on the platform? If there was a solid answer to this I suppose it could mitigate (2), but it's hard for me to believe (unless it's something very labor intensive that would make the automated nature of the system less useful)
Overall I still enjoyed the video and found parts completely hilarious (and far too realistic given my own experience on phone trees). But the above does give me pause about unreserved support for what he's doing","created_at":"2023-11-03T20:31:50Z","created_at_i":1699043510,"objectID":"38134595","parent_id":38130418,"story_id":38130418,"story_title":"I trapped scammers in an impossible maze [video]","story_url":"https://www.youtube.com/watch?v=dWzz3NeDz3E","updated_at":"2024-09-20T15:35:11Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"sublinear"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"> In fact I\u2019d love to pit a CLI veteran against a UI veteran in a contrived gauntlet of git challenges and see who reaches the finish line first. You could also determine the speed of light for each approach i.e. the minimum number of keypresses required to perform the action and then see which approach wins. Even if you had a thousand aliases, I still think a keyboard-centric UI (with good git API coverage) would win.
If they're faster than CLI it's because they are also CLI experts who have an ideological mission. Nobody cares about this situation. Overall the CLI is still faster in the majority of cases at most workplaces."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Lazygit Turns 5: Musings on Git, TUIs, and open source"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://jesseduffield.com/Lazygit-5-Years-On/"}},"_tags":["comment","author_sublinear","story_37009879"],"author":"sublinear","comment_text":"> In fact I\u2019d love to pit a CLI veteran against a UI veteran in a contrived gauntlet of git challenges and see who reaches the finish line first. You could also determine the speed of light for each approach i.e. the minimum number of keypresses required to perform the action and then see which approach wins. Even if you had a thousand aliases, I still think a keyboard-centric UI (with good git API coverage) would win.
If they're faster than CLI it's because they are also CLI experts who have an ideological mission. Nobody cares about this situation. Overall the CLI is still faster in the majority of cases at most workplaces.","created_at":"2023-08-06T06:57:35Z","created_at_i":1691305055,"objectID":"37019596","parent_id":37009879,"story_id":37009879,"story_title":"Lazygit Turns 5: Musings on Git, TUIs, and open source","story_url":"https://jesseduffield.com/Lazygit-5-Years-On/","updated_at":"2024-09-20T14:46:53Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"benreesman"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"I find this stuff fascinating, and this article is way above average for online posts about proprietary/algorithmic/quantitative/low-latency trading (very leaky Venn diagram there). I have a few nitpicks but overall it's informative and it's an interesting format: viewing an industry through the lens of a particular firm, especially one as fascinating as Jane. Anything that develops literacy in modern finance amongst the lay public is a good thing in my book.
If this stuff floats your boat I'd also recommend any of Carl Cook's talks, e.g. https://www.youtube.com/watch?v=NH1Tta7purM. Optiver is AFAIK in a somewhat different business than Jane, but they're also players (or were last I had any inside baseball).
Too many people got their worldview on this industry from "Flash Boys", and I say this as a Lewis fan, is criminally stupid at best and in bad faith at worst (if you want a well-researched, accessible alternative: https://www.amazon.com/Trading-Speed-Light-Algorithms-Transf... is about a zillion times better).
It's a pretty short list of places I'd ever go through some grueling and semi-arbitrary gauntlet to work for, but Jane is on it for sure.
I hope the author(s) do Medallion next."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Understanding Jane Street"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.thediff.co/p/jane-street"}},"_tags":["comment","author_benreesman","story_32314623"],"author":"benreesman","children":[32315475,32315662,32316011,32316024,32317036,32317360,32319472,32321943,32324232],"comment_text":"I find this stuff fascinating, and this article is way above average for online posts about proprietary/algorithmic/quantitative/low-latency trading (very leaky Venn diagram there). I have a few nitpicks but overall it's informative and it's an interesting format: viewing an industry through the lens of a particular firm, especially one as fascinating as Jane. Anything that develops literacy in modern finance amongst the lay public is a good thing in my book.
If this stuff floats your boat I'd also recommend any of Carl Cook's talks, e.g. https://www.youtube.com/watch?v=NH1Tta7purM. Optiver is AFAIK in a somewhat different business than Jane, but they're also players (or were last I had any inside baseball).
Too many people got their worldview on this industry from "Flash Boys", and I say this as a Lewis fan, is criminally stupid at best and in bad faith at worst (if you want a well-researched, accessible alternative: https://www.amazon.com/Trading-Speed-Light-Algorithms-Transf... is about a zillion times better).
It's a pretty short list of places I'd ever go through some grueling and semi-arbitrary gauntlet to work for, but Jane is on it for sure.
I hope the author(s) do Medallion next.","created_at":"2022-08-02T05:31:46Z","created_at_i":1659418306,"objectID":"32315441","parent_id":32314623,"story_id":32314623,"story_title":"Understanding Jane Street","story_url":"https://www.thediff.co/p/jane-street","updated_at":"2026-04-26T18:40:08Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"magicloop"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ghuntley","overalls"],"value":"SV companies are heavily loaded up with engineers that passed leetcode style tests. Not everyone though, due to acquisitions - engineers get grandfathered in. Leetcode coders implicitly will bias to hire other leetcode coders.\nThere will be non-SWE departments, Program Managers etc. who might be hired via the MBA-tribe. So in the end, you end up with a mix of skills (or a mix of stereotypes) in a typical SV company. I think if it were strictly just leetcoder types the company would be a disaster (but I have no data or examples - I am interested to hear about if other folks know examples).
This model of operation feels kooky but it does work on various levels (but fails in some also e.g. as panglossian fantasies ). The overall hodgepodge results in SV dominance we see today in many sectors.
Leetcode is self-fulfilling now. It's like a ticket on to the prestige and mega-bucks ride a lot of engineers desire. People pay for the ticket with 3-4 months study. Those that cannot make the time, say due to family circumstances, are actually part of the implicit unfair bias in the system.
I think what works best is when Leetcode is part of the gauntlet but not the whole game. A genuinely good person, in terms of character, together with good programming skills is what I've seen as a company's best asset. An open ended technical focussed discussion on specifically what the engineer has done lifts the lid on that part of things."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: What to do about \u2018Good at programming Bad at Leetcode\u2019"}},"_tags":["comment","author_magicloop","story_31450713"],"author":"magicloop","comment_text":"SV companies are heavily loaded up with engineers that passed leetcode style tests. Not everyone though, due to acquisitions - engineers get grandfathered in. Leetcode coders implicitly will bias to hire other leetcode coders.\nThere will be non-SWE departments, Program Managers etc. who might be hired via the MBA-tribe. So in the end, you end up with a mix of skills (or a mix of stereotypes) in a typical SV company. I think if it were strictly just leetcoder types the company would be a disaster (but I have no data or examples - I am interested to hear about if other folks know examples).
This model of operation feels kooky but it does work on various levels (but fails in some also e.g. as panglossian fantasies ). The overall hodgepodge results in SV dominance we see today in many sectors.
Leetcode is self-fulfilling now. It's like a ticket on to the prestige and mega-bucks ride a lot of engineers desire. People pay for the ticket with 3-4 months study. Those that cannot make the time, say due to family circumstances, are actually part of the implicit unfair bias in the system.
I think what works best is when Leetcode is part of the gauntlet but not the whole game. A genuinely good person, in terms of character, together with good programming skills is what I've seen as a company's best asset. An open ended technical focussed discussion on specifically what the engineer has done lifts the lid on that part of things.","created_at":"2022-05-20T20:22:51Z","created_at_i":1653078171,"objectID":"31452083","parent_id":31450713,"story_id":31450713,"story_title":"Ask HN: What to do about \u2018Good at programming Bad at Leetcode\u2019","updated_at":"2024-09-20T11:10:35Z"}],"hitsPerPage":20,"nbHits":37,"nbPages":2,"page":0,"params":"query=ghuntley+overalls&advancedSyntax=true&analyticsTags=backend","processingTimeMS":25,"processingTimingsMS":{"_request":{"roundTrip":18},"afterFetch":{"format":{"highlighting":2,"total":2},"merge":{"mergeLoop":{"prepareNextHit":6,"total":6},"total":6},"total":6},"fetch":{"query":13,"scanning":3,"total":17},"total":25},"query":"ghuntley overalls","serverTimeMS":28}