Some of the best coders and ops people I've ever worked with either never went or never graduated from college.
If you've been working 2 years as a full-time mobile dev, you're qualified for a full-time role. We pay our interns industry salaries, just for 3 months instead of 12. We do take a little bit of extra care to give them a good intro to the company and make sure we take them to a factory in their first (and only) 3 months and have a few intern events over the summer, but otherwise, it's 80-90% like having a software developer job. A few places will have some HR/recruiting drone 86 your resume because of no college. Many will not and most hiring managers who are/were programmers won't care.
tl;dr: Apply anyway (for full-time)."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Intern at a YC Company"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://blog.ycombinator.com/intern-at-a-yc-company-2019/"}},"_tags":["comment","author_sokoloff","story_19174814"],"author":"sokoloff","children":[19176059,19176156],"comment_text":"I don't want to be too blunt, but I think the barrier you are seeing is 25% real and 75% in your head.
Some of the best coders and ops people I've ever worked with either never went or never graduated from college.
If you've been working 2 years as a full-time mobile dev, you're qualified for a full-time role. We pay our interns industry salaries, just for 3 months instead of 12. We do take a little bit of extra care to give them a good intro to the company and make sure we take them to a factory in their first (and only) 3 months and have a few intern events over the summer, but otherwise, it's 80-90% like having a software developer job. A few places will have some HR/recruiting drone 86 your resume because of no college. Many will not and most hiring managers who are/were programmers won't care.
tl;dr: Apply anyway (for full-time).","created_at":"2019-02-15T23:11:14Z","created_at_i":1550272274,"objectID":"19175623","parent_id":19175536,"story_id":19174814,"story_title":"Intern at a YC Company","story_url":"https://blog.ycombinator.com/intern-at-a-yc-company-2019/","updated_at":"2024-09-20T03:46:34Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ramanujan"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"It is pretty interesting to list everything internet-related (from VOIP to bill payments to search) and realize that we're in a period similar to the era in which electricity became a utility. In the early 20th century, electricity was still uncommon in many places; Lenin's slogan[1] in the 1920s was that \"Communism = Soviet Power + Electrification of the whole country\".
Paul David[2] is an economic historian who pointed out that factories were completely redesigned once people realized they didn't need to be centered on a steam engine:
Electric light bulbs were available by 1879, and there \n were generating stations in New York and London by 1881. \n Yet a thoughtful observer in 1900 would have found little \n evidence that the \"electricity revolution\" was making \n business more efficient.\n\n Steam-powered manufacturing had linked an entire \n production line to a single huge steam engine. As a \n result, factories were stacked on many floors around the \n central engine, with drive belts all running at the same \n speed. The flow of work around the factory was governed by \n the need to put certain machines close to the steam \n engine, rather than the logic of moving the product from \n one machine to the next. When electric dynamos were first \n introduced, the steam engine would be ripped out and the \n dynamo would replace it. Productivity barely improved.\n\n Eventually, businesses figured out that factories could be \n completely redesigned on a single floor. Production lines \n were arranged to enable the smooth flow of materials \n around the factory. Most importantly, each worker could \n have his or her own little electric motor, starting it or \n stopping it at will. The improvements weren't just \n architectural but social: Once the technology allowed \n workers to make more decisions, they needed more training \n and different contracts to encourage them to take \n responsibility.\nThe overall phenomenon is really just Marc Andreessen's thesis[3] in another guise. Today, \"Software is Eating the World\", and software engineers + the internet are automating processes and arguably obsolescing many jobs. But if we warped back 80-90 years or so, we'd say that \"Electricity is Eating the World\", and electrical and mechanical engineers were the ones automating processes and obsolescing jobs.
[1] http://en.wikipedia.org/wiki/GOELRO_plan
[2] http://www.slate.com/articles/arts/the_undercover_economist/...
[3] http://online.wsj.com/article/SB1000142405311190348090457651..."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"I'm leaving the Internet for a year"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"http://www.theverge.com/2012/4/30/2988798/paul-miller-year-without-internet"}},"_tags":["comment","author_ramanujan","story_3910261"],"author":"ramanujan","children":[3913339,3913988],"comment_text":"It is pretty interesting to list everything internet-related (from VOIP to bill payments to search) and realize that we're in a period similar to the era in which electricity became a utility. In the early 20th century, electricity was still uncommon in many places; Lenin's slogan[1] in the 1920s was that \"Communism = Soviet Power + Electrification of the whole country\".
Paul David[2] is an economic historian who pointed out that factories were completely redesigned once people realized they didn't need to be centered on a steam engine:
Electric light bulbs were available by 1879, and there \n were generating stations in New York and London by 1881. \n Yet a thoughtful observer in 1900 would have found little \n evidence that the \"electricity revolution\" was making \n business more efficient.\n\n Steam-powered manufacturing had linked an entire \n production line to a single huge steam engine. As a \n result, factories were stacked on many floors around the \n central engine, with drive belts all running at the same \n speed. The flow of work around the factory was governed by \n the need to put certain machines close to the steam \n engine, rather than the logic of moving the product from \n one machine to the next. When electric dynamos were first \n introduced, the steam engine would be ripped out and the \n dynamo would replace it. Productivity barely improved.\n\n Eventually, businesses figured out that factories could be \n completely redesigned on a single floor. Production lines \n were arranged to enable the smooth flow of materials \n around the factory. Most importantly, each worker could \n have his or her own little electric motor, starting it or \n stopping it at will. The improvements weren't just \n architectural but social: Once the technology allowed \n workers to make more decisions, they needed more training \n and different contracts to encourage them to take \n responsibility.\nThe overall phenomenon is really just Marc Andreessen's thesis[3] in another guise. Today, \"Software is Eating the World\", and software engineers + the internet are automating processes and arguably obsolescing many jobs. But if we warped back 80-90 years or so, we'd say that \"Electricity is Eating the World\", and electrical and mechanical engineers were the ones automating processes and obsolescing jobs.
[1] http://en.wikipedia.org/wiki/GOELRO_plan
[2] http://www.slate.com/articles/arts/the_undercover_economist/...
[3] http://online.wsj.com/article/SB1000142405311190348090457651...","created_at":"2012-05-01T01:36:06Z","created_at_i":1335836166,"objectID":"3912559","parent_id":3910844,"story_id":3910261,"story_title":"I'm leaving the Internet for a year","story_url":"http://www.theverge.com/2012/4/30/2988798/paul-miller-year-without-internet","updated_at":"2024-09-19T18:31:52Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"theoriginaldave"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"I've struggled with ADHD my whole life. And based on my behavior, you'd think it was obvious. I have the inattentive flavor. When I'm "locked-in" on a task or subject, I can focus on it completely, and be completely inattentive to the rest of the world: the house might be on fire, but I've only got one more piece of code to finish.
Also, if I did manage to interrupt my fugue, it was almost impossible to finish the task: for a fossil collection that was half of my natural science semester grade, I spent a weekend walking a creek and limestone quarry and collected 500 fossils, twice the required, but after going home, I never sorted and labeled them, even though I knew what they were when I collected them. I ended up failing the class because of lack of 1 hour of cataloging.
This continued through the rest of school. I entered university with 29 credits due to testing (I can ace any test), but dropped out after two semesters because projects and self-study just didn't happen.
I got lucky in a job as a troubleshooter. Fly into a factory with a software problem, work 18 hour days to find and fix it, and leave a hero. Great for the career, but I would lose a fortune, because I would wait so long to do expense reports and pay credit cards.
This type of ignoring things that you desperately need to do, in lieu of the current easy to focus things is one of the major diagnostic distinctions between being lazy or unmotivated, and ADHD.
I was formally diagnosed with ADHD at age 42, and prescribed Adderall. And it was like shining a light on the world.
I'm 53 now and still taking Adderall.
There was definitely an Adderall honeymoon, but the lasting effects are also apparent.
I now 100% pay my bills on time, I maintain my car on schedule, I keep the house clean. My credit score went from about 500 (poor) to about 800 (excellent).
I still accrue debt and then have to work to pay it off, and lose money on interest, when had I planned I could have just paid cash.
And I'm still constantly harassed (good naturedly) by my Managers to do expense reports (they know I have an issue and accept it and are willing to work with me).
Probably with the right coaching, methods and strict oversight I could have done all these things without drugs, but I would have hated it.
I think the drugs improve my quality of life, and reduce the stress in my life immeasurably (I could probably measure it by stop using them for a year, but I'm not gonna!)
So I think the test needs to be:
1. Does ignoring the things you need to do make you miserable and feel like a failure?
2. Is your quality of life suffering?
3. Do you lose jobs,Relationships, because of these failures?
4. Does it prevent you from doing the things that would make you happy in life?"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Discuss ADHD and your use of medication"}},"_tags":["comment","author_theoriginaldave","story_40453713"],"author":"theoriginaldave","comment_text":"I've struggled with ADHD my whole life. And based on my behavior, you'd think it was obvious. I have the inattentive flavor. When I'm "locked-in" on a task or subject, I can focus on it completely, and be completely inattentive to the rest of the world: the house might be on fire, but I've only got one more piece of code to finish.
Also, if I did manage to interrupt my fugue, it was almost impossible to finish the task: for a fossil collection that was half of my natural science semester grade, I spent a weekend walking a creek and limestone quarry and collected 500 fossils, twice the required, but after going home, I never sorted and labeled them, even though I knew what they were when I collected them. I ended up failing the class because of lack of 1 hour of cataloging.
This continued through the rest of school. I entered university with 29 credits due to testing (I can ace any test), but dropped out after two semesters because projects and self-study just didn't happen.
I got lucky in a job as a troubleshooter. Fly into a factory with a software problem, work 18 hour days to find and fix it, and leave a hero. Great for the career, but I would lose a fortune, because I would wait so long to do expense reports and pay credit cards.
This type of ignoring things that you desperately need to do, in lieu of the current easy to focus things is one of the major diagnostic distinctions between being lazy or unmotivated, and ADHD.
I was formally diagnosed with ADHD at age 42, and prescribed Adderall. And it was like shining a light on the world.
I'm 53 now and still taking Adderall.
There was definitely an Adderall honeymoon, but the lasting effects are also apparent.
I now 100% pay my bills on time, I maintain my car on schedule, I keep the house clean. My credit score went from about 500 (poor) to about 800 (excellent).
I still accrue debt and then have to work to pay it off, and lose money on interest, when had I planned I could have just paid cash.
And I'm still constantly harassed (good naturedly) by my Managers to do expense reports (they know I have an issue and accept it and are willing to work with me).
Probably with the right coaching, methods and strict oversight I could have done all these things without drugs, but I would have hated it.
I think the drugs improve my quality of life, and reduce the stress in my life immeasurably (I could probably measure it by stop using them for a year, but I'm not gonna!)
So I think the test needs to be:
1. Does ignoring the things you need to do make you miserable and feel like a failure?
2. Is your quality of life suffering?
3. Do you lose jobs,Relationships, because of these failures?
4. Does it prevent you from doing the things that would make you happy in life?","created_at":"2024-05-24T13:58:51Z","created_at_i":1716559131,"objectID":"40466349","parent_id":40453713,"story_id":40453713,"story_title":"Ask HN: Discuss ADHD and your use of medication","updated_at":"2024-09-20T17:06:29Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"aa-jv"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"1977; I was with my Mum visiting one of her friends. Bored out of my mind I poked around her living room and peeked underneath a cover, expecting to see a typewriter - it was a teletype terminal with the earcup telephone/modem receptacle, a stream of z-paper folds neatly piled up like some holy sacrament, with which my Mums friend was learning computing. She taught me a few commands, which I still remember and use today .. "ls" .. "cd" .. "cc" .. and she showed me a few lines of this 'new powerful language' she was learning, C. It was the first time I saw a printf(), and the precise moment I learned the real difference betweeen a \\r and an \\n and what it meant to the paper scroll, rolling underneath endlessly.
1978; I travelled with my Dad to a distant outback town to visit one of his clients. While I waited for them to conclude business, a big box arrived, and the client interrupted my Dad to unpack it, with great enthusiasm. We stood by while the styrofoam was spread around the office in a flurry of unpacking and cardboard ripping chaos. The client swept all the paper nonsense off a spare desk onto the floor and promptly plonked down a brand new Commodore Pet. He poked at it a few times, promptly disclaimed "I have no idea what to do with this thing, I just know I need it <chuckle chuckle>..." Dad and the client left me alone for an hour, and I spent my time typing every single weird character I could, learning how the shift and Commodore keys worked, eventually drawing a weird maze on the screen, which I recognized much later on in my youth as what could have been the beginnings of an RPG-like map.
1980; I was a bit delinquent at a local posh college, supposed to be applying myself to my homework; instead, I followed my fellow classmates with excitement, at the end of the school day, to the local "Computer Age" shop, which had just opened up in our local neighborhood (Claremont, Perth, Western Australia) .. and there, touched my first Apple II, and then Atari 800, computers. I spent so much time after school in that shop, learning BASIC on both systems, that the salesman eventually gave me my first 5.4" floppy disc, to save all the work on in between visits. I still have that floppy, buried in the depths of some book I also value highly. Somewhere.
1982; My regular jaunts around the shopping district of Subiaco took me to every single newsstand in the suburb, and then the Tandy store, and then to my mates place, who had just gotten himself a C64 - which, much to my dismay, couldn't read the code off my Apple II 5.4" floppy, nor would it allow me to just type it all in and RUN it without a hefty serving of pokes. Nevertheless, for a little while (before my mate started getting games in the mail from his British cousins), he and I would spend our weekends porting BASIC listings from the endless magazine piles we'd collected, to the C64 - my first serious cross-platform porting project, an activity and set of experiences which led to MUCH greater success a few years later in my career.
1983; my Dad had finally agreed to get me my first computer, but it was not without great conditions - I had to work at least 4 hours every weekend with him, carting rubbish and bricks and detritous around various building sites he had his hands on, and .. we had to drive to Melbourne (from Perth) to get it, since it would be an opportunity for some quality family time vacation, as well as providing a considerable discount, since my Dad knew the computer importer who was bringing the machines to Australia. And so it was, after a 4 day trip across the Nullabor Plain, I was loaded up with an Oric-1 and a couple of cassette tapes, which I promptly plugged into every single cheap television set I could find on the way back home. Somewhere out there on the Nullabor, was a hotel TV with "Ready" burned into its screen. That was me. That was the end of quality family time on the Nullabor plain too - nobody in my family had the foggiest clue what I was doing to burn those Ready's into the screens.
1987; I have by now become an Assembly programmer and a budding C hacker, and knew far too much about anything to be bothered continuing with my studies, so I drop out of high school and get myself a job on the only electronics assembly production line in Western Australia, at that time, at first stringing endless marine cable harnesses, then touching up a near endless stream of clothes-iron temperature controllers as they pour through the wave solder, and then eventually I get to touch-up the box full of Australias first home-grown modem product - which didn't work, until me and the hardware guys realized there was not enough space between various components near the connectors - necessitating a redesign, a failed schedule, and an abandoned project as the designers gave up in light of the fierce competition from Singapore and Hong Kong and so on. Well, I at least got a patched-up PCB to play with, and with that I taught myself how to write a MODEM driver - first for CP/M, the only operating system available to me at the factory, and then .. when I showed other engineers in the lab that it could actually work, for MS-DOS, on the brand new PC the factory bought for the purpose of demonstrating the one Australian modem that was ever made in Perth. I got paid handsomely for that project, my first proper terminate and stay resident program ...
1988; I arrive in Los Angeles, California, my skills having sharpened enough to cut my teeth on writing an undelete program for the funky proprietary filesystem that Progress/4GL was using to bypass all the OS cruft that got in the way of performance. Astonished that I'd gotten it working (it was easy), the company offered me another project, and then another, and then another .. and after a year, with a litany of things under my belt - a multimedia engine that made fancy graphics on CGA and EGA (and eventually VGA) without requiring additional artist fart'ery, I was ready for more. Always another adventurous thing to write/compile/build/run/test/debug. And so it was, one beautiful glorious afternoon, a brand new MIPS Magnum pizzabox showed up on my desktop, far from the operations room I'd usually have to card my way into, halon emergency mask lessons and all, and suddenly .. I was a Unix/C programmer, and I had daemons to write, systems integrity to check, and a whole set of legacy terminal-based software to port. I need another terminal, but wasn't getting one until I actually wrote the terminal, on the brand new PC that I was also using, less enthusiastically. I wrote the terminal, I used it to log on to my MIPS box, I gained access to the ops room from a grand distance.
1990; BBS'es had my eyeballs pretty much glued to the screen, day and night, and from this it was a short step onto the Internet. I spend every moment of free time parsing USENET, downloading code, getting it to work on my pizzabox, sharpening and refining things, such that my boss just gave me endless challenges to complete - no longer doubting that it would happen, but rather expecting it.
1994; I participate in building the technical backbone of one of America's biggest ISP's at the time, and after a few weeks of hair-on-fire T1 configurations, modem-bank alignments, DNS and USENET server setups, I sit back, turn on the wire tap, and watch as a veritable firehose onslaught of daily new customers from all over Southern California spend 95% of their newly gained online time, downloading porn. (Oh, and I subscribe to minix-list to get me some hot Unix action on my mostly-useless 486, and as soon as Linus' little project hit funet.fi, I am hooked, line and sinker, as a Linux developer.)
2000; I'm burned out, I have a room full of SGI gear, and too much money. Motorbikes, Hollywood girlfriends, trips all over the country. I decide, its time to make synthesizers.
2001; I move to Germany to work for one of Germany's greatest synth companies, and spend my idle time in the Ruhrgebiet learning German and writing firmware loaders, USB drivers, and other such mundane things. I make a lot of cruddy synthesizer music, I learn how Germans party, I find myself far too many distractions.
...
2026; I'm sitting here in Austria, surrounded by 40 years of computing history, which still works just as it ever did over the decades. I can count my productive, shipped userbase for various big commercial projects literally in terms of millions, accrued over the decades, but all I really wanna do is hack on 80's computers and make them valuable again.
I've introduced 40,000+ folks in Vienna to the concept of retrocomputing, giving them hands-on time with machines many of them have never heard of, let alone considered worthy of attention .. a pimped out ZX Spectrum 128+, an Amstrad CPC6128, and an Oric Atmos - each with peripherals designed and built in the 21st Century, to bring each machine up to date with such things as USB storage, class drivers for mice and serial lines, and brand new display capabilities. The microcontrollers in these peripherals are many thousands of times more powerful than the original Z80/6502 CPU's in the 40 year old computers they are serving bytes to, but this doesn't matter any more. The only thing that matters is the fact that there are still brand new software titles being written for these machines every single day. Yes, the Oric Atmos gets new stuff, the Speccie and the Amstrad too of course, and .. any day now .. the Atmos will have a browser, surf the web, download software - finally catching up with its peers in terms of capabilities.
What a wonderful journey its been. One day, all of these machines will find other eyes and hands to entertain, but until then .. I hack on. And on. With dreams of the Nullabor journey still etched in my mind, a never to disappear "Ready" prompt floating in front my eyeballs, its time to read a bit more Hacker News, dig into some of the great repos' from todays preening, and build some stuff."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"I Owe My Life to the Commodore 64"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.goto10retro.com/p/i-owe-my-life-to-the-commodore-64"}},"_tags":["comment","author_aa-jv","story_48945260"],"author":"aa-jv","children":[48952512],"comment_text":"1977; I was with my Mum visiting one of her friends. Bored out of my mind I poked around her living room and peeked underneath a cover, expecting to see a typewriter - it was a teletype terminal with the earcup telephone/modem receptacle, a stream of z-paper folds neatly piled up like some holy sacrament, with which my Mums friend was learning computing. She taught me a few commands, which I still remember and use today .. "ls" .. "cd" .. "cc" .. and she showed me a few lines of this 'new powerful language' she was learning, C. It was the first time I saw a printf(), and the precise moment I learned the real difference betweeen a \\r and an \\n and what it meant to the paper scroll, rolling underneath endlessly.
1978; I travelled with my Dad to a distant outback town to visit one of his clients. While I waited for them to conclude business, a big box arrived, and the client interrupted my Dad to unpack it, with great enthusiasm. We stood by while the styrofoam was spread around the office in a flurry of unpacking and cardboard ripping chaos. The client swept all the paper nonsense off a spare desk onto the floor and promptly plonked down a brand new Commodore Pet. He poked at it a few times, promptly disclaimed "I have no idea what to do with this thing, I just know I need it <chuckle chuckle>..." Dad and the client left me alone for an hour, and I spent my time typing every single weird character I could, learning how the shift and Commodore keys worked, eventually drawing a weird maze on the screen, which I recognized much later on in my youth as what could have been the beginnings of an RPG-like map.
1980; I was a bit delinquent at a local posh college, supposed to be applying myself to my homework; instead, I followed my fellow classmates with excitement, at the end of the school day, to the local "Computer Age" shop, which had just opened up in our local neighborhood (Claremont, Perth, Western Australia) .. and there, touched my first Apple II, and then Atari 800, computers. I spent so much time after school in that shop, learning BASIC on both systems, that the salesman eventually gave me my first 5.4" floppy disc, to save all the work on in between visits. I still have that floppy, buried in the depths of some book I also value highly. Somewhere.
1982; My regular jaunts around the shopping district of Subiaco took me to every single newsstand in the suburb, and then the Tandy store, and then to my mates place, who had just gotten himself a C64 - which, much to my dismay, couldn't read the code off my Apple II 5.4" floppy, nor would it allow me to just type it all in and RUN it without a hefty serving of pokes. Nevertheless, for a little while (before my mate started getting games in the mail from his British cousins), he and I would spend our weekends porting BASIC listings from the endless magazine piles we'd collected, to the C64 - my first serious cross-platform porting project, an activity and set of experiences which led to MUCH greater success a few years later in my career.
1983; my Dad had finally agreed to get me my first computer, but it was not without great conditions - I had to work at least 4 hours every weekend with him, carting rubbish and bricks and detritous around various building sites he had his hands on, and .. we had to drive to Melbourne (from Perth) to get it, since it would be an opportunity for some quality family time vacation, as well as providing a considerable discount, since my Dad knew the computer importer who was bringing the machines to Australia. And so it was, after a 4 day trip across the Nullabor Plain, I was loaded up with an Oric-1 and a couple of cassette tapes, which I promptly plugged into every single cheap television set I could find on the way back home. Somewhere out there on the Nullabor, was a hotel TV with "Ready" burned into its screen. That was me. That was the end of quality family time on the Nullabor plain too - nobody in my family had the foggiest clue what I was doing to burn those Ready's into the screens.
1987; I have by now become an Assembly programmer and a budding C hacker, and knew far too much about anything to be bothered continuing with my studies, so I drop out of high school and get myself a job on the only electronics assembly production line in Western Australia, at that time, at first stringing endless marine cable harnesses, then touching up a near endless stream of clothes-iron temperature controllers as they pour through the wave solder, and then eventually I get to touch-up the box full of Australias first home-grown modem product - which didn't work, until me and the hardware guys realized there was not enough space between various components near the connectors - necessitating a redesign, a failed schedule, and an abandoned project as the designers gave up in light of the fierce competition from Singapore and Hong Kong and so on. Well, I at least got a patched-up PCB to play with, and with that I taught myself how to write a MODEM driver - first for CP/M, the only operating system available to me at the factory, and then .. when I showed other engineers in the lab that it could actually work, for MS-DOS, on the brand new PC the factory bought for the purpose of demonstrating the one Australian modem that was ever made in Perth. I got paid handsomely for that project, my first proper terminate and stay resident program ...
1988; I arrive in Los Angeles, California, my skills having sharpened enough to cut my teeth on writing an undelete program for the funky proprietary filesystem that Progress/4GL was using to bypass all the OS cruft that got in the way of performance. Astonished that I'd gotten it working (it was easy), the company offered me another project, and then another, and then another .. and after a year, with a litany of things under my belt - a multimedia engine that made fancy graphics on CGA and EGA (and eventually VGA) without requiring additional artist fart'ery, I was ready for more. Always another adventurous thing to write/compile/build/run/test/debug. And so it was, one beautiful glorious afternoon, a brand new MIPS Magnum pizzabox showed up on my desktop, far from the operations room I'd usually have to card my way into, halon emergency mask lessons and all, and suddenly .. I was a Unix/C programmer, and I had daemons to write, systems integrity to check, and a whole set of legacy terminal-based software to port. I need another terminal, but wasn't getting one until I actually wrote the terminal, on the brand new PC that I was also using, less enthusiastically. I wrote the terminal, I used it to log on to my MIPS box, I gained access to the ops room from a grand distance.
1990; BBS'es had my eyeballs pretty much glued to the screen, day and night, and from this it was a short step onto the Internet. I spend every moment of free time parsing USENET, downloading code, getting it to work on my pizzabox, sharpening and refining things, such that my boss just gave me endless challenges to complete - no longer doubting that it would happen, but rather expecting it.
1994; I participate in building the technical backbone of one of America's biggest ISP's at the time, and after a few weeks of hair-on-fire T1 configurations, modem-bank alignments, DNS and USENET server setups, I sit back, turn on the wire tap, and watch as a veritable firehose onslaught of daily new customers from all over Southern California spend 95% of their newly gained online time, downloading porn. (Oh, and I subscribe to minix-list to get me some hot Unix action on my mostly-useless 486, and as soon as Linus' little project hit funet.fi, I am hooked, line and sinker, as a Linux developer.)
2000; I'm burned out, I have a room full of SGI gear, and too much money. Motorbikes, Hollywood girlfriends, trips all over the country. I decide, its time to make synthesizers.
2001; I move to Germany to work for one of Germany's greatest synth companies, and spend my idle time in the Ruhrgebiet learning German and writing firmware loaders, USB drivers, and other such mundane things. I make a lot of cruddy synthesizer music, I learn how Germans party, I find myself far too many distractions.
...
2026; I'm sitting here in Austria, surrounded by 40 years of computing history, which still works just as it ever did over the decades. I can count my productive, shipped userbase for various big commercial projects literally in terms of millions, accrued over the decades, but all I really wanna do is hack on 80's computers and make them valuable again.
I've introduced 40,000+ folks in Vienna to the concept of retrocomputing, giving them hands-on time with machines many of them have never heard of, let alone considered worthy of attention .. a pimped out ZX Spectrum 128+, an Amstrad CPC6128, and an Oric Atmos - each with peripherals designed and built in the 21st Century, to bring each machine up to date with such things as USB storage, class drivers for mice and serial lines, and brand new display capabilities. The microcontrollers in these peripherals are many thousands of times more powerful than the original Z80/6502 CPU's in the 40 year old computers they are serving bytes to, but this doesn't matter any more. The only thing that matters is the fact that there are still brand new software titles being written for these machines every single day. Yes, the Oric Atmos gets new stuff, the Speccie and the Amstrad too of course, and .. any day now .. the Atmos will have a browser, surf the web, download software - finally catching up with its peers in terms of capabilities.
What a wonderful journey its been. One day, all of these machines will find other eyes and hands to entertain, but until then .. I hack on. And on. With dreams of the Nullabor journey still etched in my mind, a never to disappear "Ready" prompt floating in front my eyeballs, its time to read a bit more Hacker News, dig into some of the great repos' from todays preening, and build some stuff.","created_at":"2026-07-17T17:23:22Z","created_at_i":1784309002,"objectID":"48949868","parent_id":48945260,"story_id":48945260,"story_title":"I Owe My Life to the Commodore 64","story_url":"https://www.goto10retro.com/p/i-owe-my-life-to-the-commodore-64","updated_at":"2026-07-19T00:40:35Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ben_w"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"> How long do you think it will be before a reliable dish washing / laundry doing domestic bot is on the market? a decade or two?
Depends if they only do this, or if they're fully general robots that also do this.
The former, remember we've already got dishwashers and washing machines.
The latter, to a standard that won't lead to regular news stories about how it tries to put a pet corgi into the bread bin or fry someone's pet goldfish for dinner, the problem space they need to navigate (literally and metaphorically) is much more complex than what is needed by level-5 FSD. Right now level-5 isn't a solved problem (running into parked cars should be "never happens" not merely "rare"), suggesting that even the more power-intensive compute (800 watts) isn't sufficient for the easier challenge of level-5 FSD, let alone domestic robots, and at that power level such domestic robots would need to be sold with an external computer running of mains not an internal one that runs of batteries.
Even if the software is a solved problem, which it isn't, getting the compute into a mobile body with the much smaller power envelope of an android vs. a car means the gap between level-5 and such robots is order of 5-10 years. But, to emphasise, this countdown has not yet started, and I'm not saying it will be between 2030-35.
For large-scale use, the global electrical power supply is currently 375 W/capita, and even with my expectation of wind+PV allowing us to double this by 2032, there's not enough electrical capacity for a huge number of these robots in this form on that timescale.
Regarding clanking self replication, that depends on the exact details of what you mean. It took a while to get the taxonomy of level-[12345] self driving, and right now we're too far from meaningful self replication to even make a taxonomy of self-replication:
\u2022 If all you need is industrial robot arms that work in a factory that makes more industrial robot arms, we've probably already got that.
\u2022 If you need the robots to be sufficiently general that we can fill up the payload bay of Starship with ~1000 of them, dump them all on the moon, and a few months later they've bashed enough rocks together to make an ore processing system, and a year later they've built a factory that makes more of themselves, that's much much harder and I don't think we've got enough information yet to even enumerate all the steps between here and there."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"UniTree Humanoid Companion"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.youtube.com/watch?v=v1Q4Su54iho"}},"_tags":["comment","author_ben_w","story_44688166"],"author":"ben_w","comment_text":"> How long do you think it will be before a reliable dish washing / laundry doing domestic bot is on the market? a decade or two?
Depends if they only do this, or if they're fully general robots that also do this.
The former, remember we've already got dishwashers and washing machines.
The latter, to a standard that won't lead to regular news stories about how it tries to put a pet corgi into the bread bin or fry someone's pet goldfish for dinner, the problem space they need to navigate (literally and metaphorically) is much more complex than what is needed by level-5 FSD. Right now level-5 isn't a solved problem (running into parked cars should be "never happens" not merely "rare"), suggesting that even the more power-intensive compute (800 watts) isn't sufficient for the easier challenge of level-5 FSD, let alone domestic robots, and at that power level such domestic robots would need to be sold with an external computer running of mains not an internal one that runs of batteries.
Even if the software is a solved problem, which it isn't, getting the compute into a mobile body with the much smaller power envelope of an android vs. a car means the gap between level-5 and such robots is order of 5-10 years. But, to emphasise, this countdown has not yet started, and I'm not saying it will be between 2030-35.
For large-scale use, the global electrical power supply is currently 375 W/capita, and even with my expectation of wind+PV allowing us to double this by 2032, there's not enough electrical capacity for a huge number of these robots in this form on that timescale.
Regarding clanking self replication, that depends on the exact details of what you mean. It took a while to get the taxonomy of level-[12345] self driving, and right now we're too far from meaningful self replication to even make a taxonomy of self-replication:
\u2022 If all you need is industrial robot arms that work in a factory that makes more industrial robot arms, we've probably already got that.
\u2022 If you need the robots to be sufficiently general that we can fill up the payload bay of Starship with ~1000 of them, dump them all on the moon, and a few months later they've bashed enough rocks together to make an ore processing system, and a year later they've built a factory that makes more of themselves, that's much much harder and I don't think we've got enough information yet to even enumerate all the steps between here and there.","created_at":"2025-07-26T10:22:34Z","created_at_i":1753525354,"objectID":"44692935","parent_id":44689163,"story_id":44688166,"story_title":"UniTree Humanoid Companion","story_url":"https://www.youtube.com/watch?v=v1Q4Su54iho","updated_at":"2025-07-26T10:25:09Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"Quothling"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"Working in an investment bank that's build around 800 solar plants over the world in the previous decade with steep competition from a lot of other investors I'd argue that not everyone missed it. The next big boom is in battery storage, but the projects and investments took off a year or two ago with production factories nearing completion. Part of what shot it up was when various EU countries raised their green energy targets from 60-80%. In Germany alone that will require a 800% growth for battery storage financials before 2030.
I'm not sure which semi-professional to professional investors missed that here in Europe though. In fact we've seen so many companies shoot up around these things that the recent economic downturns is causing quite a lot of bankruptcies. Which are then bought up by larger investment funds.
From a tech perspective... There is going to be a lot of potential in building the technical layer between a plant and the "internet". None of the current systems, including market leaders that I'm not going to name drop, aren't very good. Part of the reason is because things like inverters and other solar plant tech has absolutely no standardization. They are ridiculously different, meaning you need actual programmers to onboard each plant unless someone builds some form of engineering standard. Currently it seems like each piece of tech basically has it's software designed by the individual engineer who made it. What is even more hilarious is that the systems made for solar, wind, batteries and so on don't work together."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Why everyone missed solar's exponential growth"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.exponentialview.co/p/the-forecasters-gap"}},"_tags":["comment","author_Quothling","story_42200149"],"author":"Quothling","children":[42201841,42202080,42202215,42204364],"comment_text":"Working in an investment bank that's build around 800 solar plants over the world in the previous decade with steep competition from a lot of other investors I'd argue that not everyone missed it. The next big boom is in battery storage, but the projects and investments took off a year or two ago with production factories nearing completion. Part of what shot it up was when various EU countries raised their green energy targets from 60-80%. In Germany alone that will require a 800% growth for battery storage financials before 2030.
I'm not sure which semi-professional to professional investors missed that here in Europe though. In fact we've seen so many companies shoot up around these things that the recent economic downturns is causing quite a lot of bankruptcies. Which are then bought up by larger investment funds.
From a tech perspective... There is going to be a lot of potential in building the technical layer between a plant and the "internet". None of the current systems, including market leaders that I'm not going to name drop, aren't very good. Part of the reason is because things like inverters and other solar plant tech has absolutely no standardization. They are ridiculously different, meaning you need actual programmers to onboard each plant unless someone builds some form of engineering standard. Currently it seems like each piece of tech basically has it's software designed by the individual engineer who made it. What is even more hilarious is that the systems made for solar, wind, batteries and so on don't work together.","created_at":"2024-11-21T06:49:18Z","created_at_i":1732171758,"objectID":"42201788","parent_id":42200149,"story_id":42200149,"story_title":"Why everyone missed solar's exponential growth","story_url":"https://www.exponentialview.co/p/the-forecasters-gap","updated_at":"2024-11-29T12:52:28Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"beart"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"This is the path I took when purchasing my most recent vehicle. A ford with a ~5" non-touch screen. Basically just big enough to show you the backup camera and the radio station.
I love having physical controls for everything.
Honestly though, as much as I hate touch screens, this thing is terrible.
The biggest issue is the software, which admittedly has nothing to do with the form factor. It just crashes all the time, and is annoying to use. The backup camera is difficult to see. Not having navigation readily available is frustrating once you've driven vehicles that do have it. Mounting your phone is not a good substitute.
I think you really have to shoot for a design where 80-90% of your functions are physical buttons/knobs while also having the advanced software and high definition display that android auto/car play offer."},"story_title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["software"],"value":"GM to Cut More Than 1k Software Engineers, Mostly in US"},"story_url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["software"],"value":"https://webcache.googleusercontent.com/search?q=cache:https://www.bloomberg.com/news/articles/2024-08-19/gm-to-cut-more-than-1-000-software-engineers-mostly-in-the-us"}},"_tags":["comment","author_beart","story_41296152"],"author":"beart","children":[41297226,41297234],"comment_text":"This is the path I took when purchasing my most recent vehicle. A ford with a ~5" non-touch screen. Basically just big enough to show you the backup camera and the radio station.
I love having physical controls for everything.
Honestly though, as much as I hate touch screens, this thing is terrible.
The biggest issue is the software, which admittedly has nothing to do with the form factor. It just crashes all the time, and is annoying to use. The backup camera is difficult to see. Not having navigation readily available is frustrating once you've driven vehicles that do have it. Mounting your phone is not a good substitute.
I think you really have to shoot for a design where 80-90% of your functions are physical buttons/knobs while also having the advanced software and high definition display that android auto/car play offer.","created_at":"2024-08-20T04:19:17Z","created_at_i":1724127557,"objectID":"41296751","parent_id":41296591,"story_id":41296152,"story_title":"GM to Cut More Than 1k Software Engineers, Mostly in US","story_url":"https://webcache.googleusercontent.com/search?q=cache:https://www.bloomberg.com/news/articles/2024-08-19/gm-to-cut-more-than-1-000-software-engineers-mostly-in-the-us","updated_at":"2024-09-20T17:43:29Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"alephnerd"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"> But when you go fishing for cheap, which is usually the reason for seeking it in the first place, you're never going to find it
That's the thing - employers aren't paying peanuts abroad anymore either.
Just take a look at YC salaries in India, Czechia, or China (the level of talent that FAANG in those countries would hire at).
Base salary plus bonus is at most 50% less than in the US, so factoring taxes and legal fees you save at best 40%, which isn't significant in an industry with 80-90% profit margins.
It's literally about talent at the end of the day - it's just easier to scale out entirely new teams to speed up delivering multiple features at once.
If you pay peanuts, then software is just a cost center for you (and for most industries, it is)"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Move over, remote jobs. CEOs say borderless talent is the future of tech work"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.cnbc.com/2024/06/30/move-over-remote-ceos-say-borderless-talent-future-tech-jobs.html"}},"_tags":["comment","author_alephnerd","story_40891312"],"author":"alephnerd","comment_text":"> But when you go fishing for cheap, which is usually the reason for seeking it in the first place, you're never going to find it
That's the thing - employers aren't paying peanuts abroad anymore either.
Just take a look at YC salaries in India, Czechia, or China (the level of talent that FAANG in those countries would hire at).
Base salary plus bonus is at most 50% less than in the US, so factoring taxes and legal fees you save at best 40%, which isn't significant in an industry with 80-90% profit margins.
It's literally about talent at the end of the day - it's just easier to scale out entirely new teams to speed up delivering multiple features at once.
If you pay peanuts, then software is just a cost center for you (and for most industries, it is)","created_at":"2024-07-06T18:57:33Z","created_at_i":1720292253,"objectID":"40892407","parent_id":40892044,"story_id":40891312,"story_title":"Move over, remote jobs. CEOs say borderless talent is the future of tech work","story_url":"https://www.cnbc.com/2024/06/30/move-over-remote-ceos-say-borderless-talent-future-tech-jobs.html","updated_at":"2024-09-20T17:27:39Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"TimBuckToot"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"The Amiga was like a dream machine when it hit the market, most people I knew had C64's at the time and the Amiga was just light years ahead in terms of capabilities. A few people I knew had PC's and the Amiga absolutely left them in the dust for years, it only really seemed to catch up around 1993 once Doom hit and VGA / SVGA cards became somewhat more affordable, that was roughly eight years after the Amiga was released.
All of that being said, I lived in a part of the world in the 80/90's where PC tech was very expensive at the time, and the humble Amiga was very reasonably priced in comparison, it had just enough wow factor initially to make you believe, with lots of custom chips to support that ambition. It was the sort of computer that was able to crossover into every space mostly. If you lived in the US, you were likely trying to decide between a Tandy PC and Apple or a console which were way more affordable on home soil and already had a large catalogue of titles and mainstream developer support and business use (for the PC).
Where I think things went wrong for the Amiga, aside from Irving Gould and his rather hopeless management squad, was that the Amiga was marketed badly and poorly positioned in the US, with Commodore having a somewhat sketchy relationship with smaller computers stores across the country which almost guaranteed they got a bit less sunshine compared to other devices in store (thanks in part to Jack Tramiel and his aggressive price cutting tactics with the C64 in the early 80's). Europe, Australia, and South Africa all had brilliant marketing and distribution in comparison, and it was thanks to these markets the machine endured as long as it did.
The Amiga however ended up fighting too many battles on too many fronts and had very poor strategic direction in general. They needed R&D badly and a decent amount of it to maintain a hardware advantage at a low cost price point, instead they got rid of many of their great hardware engineers and specialists when their initial project had finished so they could save a buck and ultimately seeded their talent to the competition...
They should've had a VGA or better machine out to market by 1990, which they didn't and that was a major loss in terms of the hardware game. PAL and NTSC Amiga machines were also different beasts, with the PAL machines coming out on top in terms of resolution, compatibility and with more developer support in general.
But it's real losing hands (and there were a few) were: \n1. Too many frequent revisions of it's main product which were superseded shortly after - Amiga 500+ or Amiga 600 anyone? \n2. No cartridge or card slot while fighting a battle against consoles for market share and not having the foresight to build a low cost standardized adapter that could've been used as one (via the almost never used SCSI connector). \n3. Releasing the CD32 Console, but not releasing an Amiga 1200 CD drive or the same capabilities across their all their home desktops (i.e. Akiko chip and CD32 firmware). \n4. Releasing an upgraded version of Workbench / bios which killed compatibility with a third of Amiga's back catalogue of software titles (that was bright). \n5. Having different resolutions and cpu speeds between countries for the same product - believe it or not there's a way to separate voltage speeds from the CPU frequency, and to have resolutions that can be complimentary across PAL and NTSC regions (just ask Nintendo and Sega), and even more amazingly if you just add a spot of frame skipping into your NTSC title it can work just fine in PAL at full speed. \n6. RGB 15khz is great, but not properly implemented to support higher resolutions, so everything looked a super flickery when you went high res - it was basically a feature you couldn't use or had to suffer through if you did use it (Until the Amber Chip appeared on the A3000 / A4000 range).\n7. The HAM mode was revolutionary but was extremely difficult to use except for the odd digitized image - another feature that people couldn't use, and wasn't made easier to use... \n8. Chip ram should've been the only RAM in the system and had a faster bus speed to support normal transport speeds across memory. It was bad because it was very limited.\n9. No upgrades across the sound channels or quality / frequency. \nand lastly... \n10. Commodore's main bread butter with the Amiga was with the desktop hobbyist home user - this segment should've been it's no 1 priority and been given first class treatment all the way. The UK and German offices understood this, but not in the US with the final nail in the coffin being the poor launch sales of the CD32 in the US (Went great in the UK). This was supported by a lack of killer titles upon release such as (Street Fighter 2, Mortal Kombat) and the general apathy to the Commodore brand in the US overall due to years of neglect and bad decisions.
The beautiful Amiga died because of idiocy and neglect - Commodore was doomed to die, it was just a rotten shame that the Amiga computer line went with it."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Amiga's final days as chronicled by the Village Voice 30 years ago"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.amigalove.com/viewtopic.php?t=2697"}},"_tags":["comment","author_TimBuckToot","story_39583492"],"author":"TimBuckToot","comment_text":"The Amiga was like a dream machine when it hit the market, most people I knew had C64's at the time and the Amiga was just light years ahead in terms of capabilities. A few people I knew had PC's and the Amiga absolutely left them in the dust for years, it only really seemed to catch up around 1993 once Doom hit and VGA / SVGA cards became somewhat more affordable, that was roughly eight years after the Amiga was released.
All of that being said, I lived in a part of the world in the 80/90's where PC tech was very expensive at the time, and the humble Amiga was very reasonably priced in comparison, it had just enough wow factor initially to make you believe, with lots of custom chips to support that ambition. It was the sort of computer that was able to crossover into every space mostly. If you lived in the US, you were likely trying to decide between a Tandy PC and Apple or a console which were way more affordable on home soil and already had a large catalogue of titles and mainstream developer support and business use (for the PC).
Where I think things went wrong for the Amiga, aside from Irving Gould and his rather hopeless management squad, was that the Amiga was marketed badly and poorly positioned in the US, with Commodore having a somewhat sketchy relationship with smaller computers stores across the country which almost guaranteed they got a bit less sunshine compared to other devices in store (thanks in part to Jack Tramiel and his aggressive price cutting tactics with the C64 in the early 80's). Europe, Australia, and South Africa all had brilliant marketing and distribution in comparison, and it was thanks to these markets the machine endured as long as it did.
The Amiga however ended up fighting too many battles on too many fronts and had very poor strategic direction in general. They needed R&D badly and a decent amount of it to maintain a hardware advantage at a low cost price point, instead they got rid of many of their great hardware engineers and specialists when their initial project had finished so they could save a buck and ultimately seeded their talent to the competition...
They should've had a VGA or better machine out to market by 1990, which they didn't and that was a major loss in terms of the hardware game. PAL and NTSC Amiga machines were also different beasts, with the PAL machines coming out on top in terms of resolution, compatibility and with more developer support in general.
But it's real losing hands (and there were a few) were: \n1. Too many frequent revisions of it's main product which were superseded shortly after - Amiga 500+ or Amiga 600 anyone? \n2. No cartridge or card slot while fighting a battle against consoles for market share and not having the foresight to build a low cost standardized adapter that could've been used as one (via the almost never used SCSI connector). \n3. Releasing the CD32 Console, but not releasing an Amiga 1200 CD drive or the same capabilities across their all their home desktops (i.e. Akiko chip and CD32 firmware). \n4. Releasing an upgraded version of Workbench / bios which killed compatibility with a third of Amiga's back catalogue of software titles (that was bright). \n5. Having different resolutions and cpu speeds between countries for the same product - believe it or not there's a way to separate voltage speeds from the CPU frequency, and to have resolutions that can be complimentary across PAL and NTSC regions (just ask Nintendo and Sega), and even more amazingly if you just add a spot of frame skipping into your NTSC title it can work just fine in PAL at full speed. \n6. RGB 15khz is great, but not properly implemented to support higher resolutions, so everything looked a super flickery when you went high res - it was basically a feature you couldn't use or had to suffer through if you did use it (Until the Amber Chip appeared on the A3000 / A4000 range).\n7. The HAM mode was revolutionary but was extremely difficult to use except for the odd digitized image - another feature that people couldn't use, and wasn't made easier to use... \n8. Chip ram should've been the only RAM in the system and had a faster bus speed to support normal transport speeds across memory. It was bad because it was very limited.\n9. No upgrades across the sound channels or quality / frequency. \nand lastly... \n10. Commodore's main bread butter with the Amiga was with the desktop hobbyist home user - this segment should've been it's no 1 priority and been given first class treatment all the way. The UK and German offices understood this, but not in the US with the final nail in the coffin being the poor launch sales of the CD32 in the US (Went great in the UK). This was supported by a lack of killer titles upon release such as (Street Fighter 2, Mortal Kombat) and the general apathy to the Commodore brand in the US overall due to years of neglect and bad decisions.
The beautiful Amiga died because of idiocy and neglect - Commodore was doomed to die, it was just a rotten shame that the Amiga computer line went with it.","created_at":"2024-03-04T15:35:15Z","created_at_i":1709566515,"objectID":"39591651","parent_id":39583492,"story_id":39583492,"story_title":"Amiga's final days as chronicled by the Village Voice 30 years ago","story_url":"https://www.amigalove.com/viewtopic.php?t=2697","updated_at":"2024-09-20T16:34:36Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"narutosasuke"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"> Who is "we"?
I'm confused why this was focused on so much. I guess "people that use DEI for hiring". Or when I was speaking of heuristics generally that could be "hiring managers in general".
> So if "we" can't be bothered to use very common-sensical criteria on what actually being "diverse" means then the DEI initiatives will keep being hijacked by opportunistic individuals (as it currently seems to be the case).
It's not controversial to say that women have largely different life experience/background than men. It's not controversial to say those of different races/cultures have largely different life experiences/background than a young white suburban software engineer (which has historically been the majority of the industry).
There are an infinite number of factors that go into a person and there will always be something missed with whatever way we slice it. "common-sensical" criteria in that world will end up as a wishy washy "we hire those with diverse opinions and backgrounds". But really that means I'm just going to hire the person I get along with best that answers the questions I happen to have, biases included.
Does it feel good that I am not focused on in DEI as a young white man? No, it sucks to be reduced to something I can't change. But I'm not arrogant enough to pretend that I'm not commonly represented at all levels of tech.
> I'd like to see exhaustive and irrefutable numerical proof if you choose to double down on this flawed assertion.
Asking for an "irrefutable numerical proof" isn't an invitation for debate or speaking from curiosity. It can be lobbed at literally any position one takes.
> It's extremely easy in tech in fact. After being an interviewer and an interviewee for 22 years, I have seen this shake out dozens of times: a round of interviews establishes with 80-90% certainty that (a) the candidate is capable and (b) matches the company's culture.
You're blowing by that I was responding to
"Some people just like to worth with really intelligent, hard working people and don't care anything else about the person"
If we're talking about company culture now, then we're talking about something different.
> Not sure why you and others are constantly trying to shift goal posts to other areas when the original article clearly qualifies its statement with the word "tech".
The comment I was responding to spoke generally so I spoke generally. I applied it to both tech and nontech with examples though."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Tech's DEI backlash is here"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://techcrunch.com/2023/12/21/techs-dei-backlash-is-here/"}},"_tags":["comment","author_narutosasuke","story_38726911"],"author":"narutosasuke","children":[38737508],"comment_text":"> Who is "we"?
I'm confused why this was focused on so much. I guess "people that use DEI for hiring". Or when I was speaking of heuristics generally that could be "hiring managers in general".
> So if "we" can't be bothered to use very common-sensical criteria on what actually being "diverse" means then the DEI initiatives will keep being hijacked by opportunistic individuals (as it currently seems to be the case).
It's not controversial to say that women have largely different life experience/background than men. It's not controversial to say those of different races/cultures have largely different life experiences/background than a young white suburban software engineer (which has historically been the majority of the industry).
There are an infinite number of factors that go into a person and there will always be something missed with whatever way we slice it. "common-sensical" criteria in that world will end up as a wishy washy "we hire those with diverse opinions and backgrounds". But really that means I'm just going to hire the person I get along with best that answers the questions I happen to have, biases included.
Does it feel good that I am not focused on in DEI as a young white man? No, it sucks to be reduced to something I can't change. But I'm not arrogant enough to pretend that I'm not commonly represented at all levels of tech.
> I'd like to see exhaustive and irrefutable numerical proof if you choose to double down on this flawed assertion.
Asking for an "irrefutable numerical proof" isn't an invitation for debate or speaking from curiosity. It can be lobbed at literally any position one takes.
> It's extremely easy in tech in fact. After being an interviewer and an interviewee for 22 years, I have seen this shake out dozens of times: a round of interviews establishes with 80-90% certainty that (a) the candidate is capable and (b) matches the company's culture.
You're blowing by that I was responding to
"Some people just like to worth with really intelligent, hard working people and don't care anything else about the person"
If we're talking about company culture now, then we're talking about something different.
> Not sure why you and others are constantly trying to shift goal posts to other areas when the original article clearly qualifies its statement with the word "tech".
The comment I was responding to spoke generally so I spoke generally. I applied it to both tech and nontech with examples though.","created_at":"2023-12-22T18:49:34Z","created_at_i":1703270974,"objectID":"38737193","parent_id":38731263,"story_id":38726911,"story_title":"Tech's DEI backlash is here","story_url":"https://techcrunch.com/2023/12/21/techs-dei-backlash-is-here/","updated_at":"2024-09-20T15:59:15Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"somenameforme"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"You run into the typical neural net problem with this logic. OpenAI (or at least Sam Altman) have already publicly acknowledged that the diminishing returns they're seeing in terms of model size are sufficient to effectively declare that 'the age of giant models is already over.' [1] It seems many people were unaware of his comments on this topic.
Neural networks in literally every other field always repeat the exact same pattern. You can get from 0-80 without breaking a sweat. 80-90 is dramatically harder but you finally get there. So everybody imagines getting from 90-100 will be little more than a matter of a bit more compute and a bit more massaging of the model. But it turns out that each fraction of a percent progress you make starts becoming exponentially more difficult - and you eventually run into an asymptote that's nowhere near what you are aiming for.
A prediction based on the typical history of neural nets would be that OpenAI will be able to continue to make progress on extremely specific metrics, like scoring well on some test or another, largely by hardcoding case-specific workarounds and tweaks. But in terms of general model usage, we're unlikely to see any real revolutionary leaps in the foreseeable future.
If we see model accuracy increase I'd expect it to be thanks not to model improvement, but instead by doing something like adding a second layer where the software cross references the generated output against a 'fact database' and regenerates its answer when some correlation factor is insufficiently high. Of course that'd completely cripple the model's ability to ever move 'beyond' its training. It'd be like if mankind was forced to double check that any response on astronomy we made confirmed that the Earth is indeed the center of the universe, with no ability to ever change that ourselves.
[1] - https://www.wired.com/story/openai-ceo-sam-altman-the-age-of..."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"StarCoder and StarCoderBase: 15.5B parameter models with 8K context length"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://arxiv.org/abs/2305.06161"}},"_tags":["comment","author_somenameforme","story_35954481"],"author":"somenameforme","children":[35958030,35959336],"comment_text":"You run into the typical neural net problem with this logic. OpenAI (or at least Sam Altman) have already publicly acknowledged that the diminishing returns they're seeing in terms of model size are sufficient to effectively declare that 'the age of giant models is already over.' [1] It seems many people were unaware of his comments on this topic.
Neural networks in literally every other field always repeat the exact same pattern. You can get from 0-80 without breaking a sweat. 80-90 is dramatically harder but you finally get there. So everybody imagines getting from 90-100 will be little more than a matter of a bit more compute and a bit more massaging of the model. But it turns out that each fraction of a percent progress you make starts becoming exponentially more difficult - and you eventually run into an asymptote that's nowhere near what you are aiming for.
A prediction based on the typical history of neural nets would be that OpenAI will be able to continue to make progress on extremely specific metrics, like scoring well on some test or another, largely by hardcoding case-specific workarounds and tweaks. But in terms of general model usage, we're unlikely to see any real revolutionary leaps in the foreseeable future.
If we see model accuracy increase I'd expect it to be thanks not to model improvement, but instead by doing something like adding a second layer where the software cross references the generated output against a 'fact database' and regenerates its answer when some correlation factor is insufficiently high. Of course that'd completely cripple the model's ability to ever move 'beyond' its training. It'd be like if mankind was forced to double check that any response on astronomy we made confirmed that the Earth is indeed the center of the universe, with no ability to ever change that ourselves.
[1] - https://www.wired.com/story/openai-ceo-sam-altman-the-age-of...","created_at":"2023-05-16T04:07:37Z","created_at_i":1684210057,"objectID":"35957652","parent_id":35956261,"story_id":35954481,"story_title":"StarCoder and StarCoderBase: 15.5B parameter models with 8K context length","story_url":"https://arxiv.org/abs/2305.06161","updated_at":"2024-09-20T14:03:34Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"loso"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"10 years ago I was a small time hip hop DJ who did some clubs and radio. Before I became a programmer, my dream was to become the next Puffy. I studied the industry big time (reading books, talking to people inside, doing interviews) and still do. I also worked security for years for the main purpose of being able to interact with the people behind the scenes (managers, promoters, etc). So even though I am not an expert, I know a decent amount about the music industry.
First of all I see some talk about how you can predict a hit. No, you can't. You have basic formulas that you know can do okay but a hit is usually a surprise. No one could predict Adele would be as big as she is now. Her first album was very good but it did not pick up steam in the US market. Lady Gaga was bounced around as well. She was almost dropped from her record label until Akon picked up the slack and was able to do something with her. Basically what I'm saying is that after something is made you may be able to tell that it is going to be big but you have to craft it first. While crafting it, you are not going by formula but what feels right. You don't know that it is going to be a hit while you are making it.
Second, you have to understand the record label model to be able to disrupt it. The point systems, mechanical royalties, publishing, different rules for selling overseas, etc. This is what the labels are good at. This is the moat to their castle. In the past, (before around 2002 or so) record labels would put out a bunch of artist and see which one stuck. They would lose money on 80 - 90% of their artist because all they needed was one or two to make it big to make their money back. Minor artist would also be able to make the label profits through the use of singles. But the record labels gave up the singles model for the larger profit margins of albums. Now it is different.
Now they have to be real careful who they put out. The profit margins are a lot thinner now than what they used to be. Piracy is one reason but a bigger reason is the more places for people to be able to spend their disposable income. The record labels did not make a pivot to change the way they did business when the economic problems hit. Instead the industry shrunk down and now we have less labels. But they are still doing business in the same fundamentally screwed up way that they did in the past. In their mind, why change something that has worked for over 50 years. Think Kodak and the reason for their recent bankruptcy.
The one thing they did change is how to make artist pay more for the squeeze that the record labels are feeling. New artist these days have to bring to the table a built in fan base. The label then will try to expand that base. For the most part, if you want to be signed to a record label these days then they have to know that you have a track record with fans. Take an artist like Whiz Khalifa. He already had a big buzz when the record labels signed him. He could have made a lot of money staying independent. But most artist want the international fame that accompanies the major record labels. So he signed with them so they could push him internationally. In the past having a buzz would help you get signed but it was not the only factor. Now in most cases, you HAVE to have a big buzz.
The labels are also making the artist pay more in a new type of deal structure. This signing is called a 360 deal. In the past, the deal basically was that an artist would receive a (relatively) low amount of points on a record put out. This is especially true if you were a new artist. But it didn't matter because if the record did well they could tour off of it. The tour and merchandise money was basically the artist only. This is how they would get rich. Now record labels have made sure they get a part of that money as well. A 360 deal makes an artist have to cut the record label in when it comes to merchandising and tour money. AND they still get low points on the album. You either accept this new industry practice or you do not get signed.
The record labels are ripe for disruption. New artist are available all of the time. But you not only have to study what the labels got wrong, you also have to study what they got right. First, musical talent is only part of the equation when it comes to selling music. People have to be personally invested in the artist. They have to like them and think that they know them. Record labels are experts at this. An artist either has to relate to the consumer or the artist has to be someone they idolize. Think of it like the Fanboy trap that we see when it comes to tech. Go to any gossip blog site. People argue about the artist as if they know them. Some people love them for no reason and others hate them for no reason. But the key factor is that they love talking about them. So if you build software to disrupt the industry, this has to be in the equation. Your service has to allow people to have a personal connection with the artist. None of the music services now do this. They let you listen to the artist but the connection is already built in from somewhere else. I would say the best right now at doing this is Youtube. But even Youtube could do a lot better job of it. MySpace was kinda on the right path but bad design decisions killed that.
Second, a new service to disrupt the industry would also have to think about the people who make the artist what they are. The stylist, the make up people, the manager, the promoters, the background singers, and etc. An artist is a package. And you would have to allow for that package to be dealt with in a new disruption model. The record labels feed a lot of people and a lot of those people are vital for a record and an artist to be pushed.
I have a lot more that I have to say about this subject but I have already built up a wall of text so I'll cut it here. I'm a little bit passionate about this topic because I have seen for years how the label system screws both artist, DJ's and fans. They need to be brought down and brought down hard but it will not be easy. The one thing that you will find is that most artist have no business sense. They want to be able to create and that's it. We see this line of thinking with some software programmers as well. But programmers usually keep a steady job. Most artist have a short shelf life. It's not until they are broke that they realize how they have been screwed. Before that happens though, they love the industry."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Can we kill the music business too?"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"http://audiosearch.blogspot.com/2012/01/can-we-kill-music-business-too.html"}},"_tags":["comment","author_loso","story_3492912"],"author":"loso","children":[3493697,3493744,3493941,3494298,3495127,3495166,3495459],"comment_text":"10 years ago I was a small time hip hop DJ who did some clubs and radio. Before I became a programmer, my dream was to become the next Puffy. I studied the industry big time (reading books, talking to people inside, doing interviews) and still do. I also worked security for years for the main purpose of being able to interact with the people behind the scenes (managers, promoters, etc). So even though I am not an expert, I know a decent amount about the music industry.
First of all I see some talk about how you can predict a hit. No, you can't. You have basic formulas that you know can do okay but a hit is usually a surprise. No one could predict Adele would be as big as she is now. Her first album was very good but it did not pick up steam in the US market. Lady Gaga was bounced around as well. She was almost dropped from her record label until Akon picked up the slack and was able to do something with her. Basically what I'm saying is that after something is made you may be able to tell that it is going to be big but you have to craft it first. While crafting it, you are not going by formula but what feels right. You don't know that it is going to be a hit while you are making it.
Second, you have to understand the record label model to be able to disrupt it. The point systems, mechanical royalties, publishing, different rules for selling overseas, etc. This is what the labels are good at. This is the moat to their castle. In the past, (before around 2002 or so) record labels would put out a bunch of artist and see which one stuck. They would lose money on 80 - 90% of their artist because all they needed was one or two to make it big to make their money back. Minor artist would also be able to make the label profits through the use of singles. But the record labels gave up the singles model for the larger profit margins of albums. Now it is different.
Now they have to be real careful who they put out. The profit margins are a lot thinner now than what they used to be. Piracy is one reason but a bigger reason is the more places for people to be able to spend their disposable income. The record labels did not make a pivot to change the way they did business when the economic problems hit. Instead the industry shrunk down and now we have less labels. But they are still doing business in the same fundamentally screwed up way that they did in the past. In their mind, why change something that has worked for over 50 years. Think Kodak and the reason for their recent bankruptcy.
The one thing they did change is how to make artist pay more for the squeeze that the record labels are feeling. New artist these days have to bring to the table a built in fan base. The label then will try to expand that base. For the most part, if you want to be signed to a record label these days then they have to know that you have a track record with fans. Take an artist like Whiz Khalifa. He already had a big buzz when the record labels signed him. He could have made a lot of money staying independent. But most artist want the international fame that accompanies the major record labels. So he signed with them so they could push him internationally. In the past having a buzz would help you get signed but it was not the only factor. Now in most cases, you HAVE to have a big buzz.
The labels are also making the artist pay more in a new type of deal structure. This signing is called a 360 deal. In the past, the deal basically was that an artist would receive a (relatively) low amount of points on a record put out. This is especially true if you were a new artist. But it didn't matter because if the record did well they could tour off of it. The tour and merchandise money was basically the artist only. This is how they would get rich. Now record labels have made sure they get a part of that money as well. A 360 deal makes an artist have to cut the record label in when it comes to merchandising and tour money. AND they still get low points on the album. You either accept this new industry practice or you do not get signed.
The record labels are ripe for disruption. New artist are available all of the time. But you not only have to study what the labels got wrong, you also have to study what they got right. First, musical talent is only part of the equation when it comes to selling music. People have to be personally invested in the artist. They have to like them and think that they know them. Record labels are experts at this. An artist either has to relate to the consumer or the artist has to be someone they idolize. Think of it like the Fanboy trap that we see when it comes to tech. Go to any gossip blog site. People argue about the artist as if they know them. Some people love them for no reason and others hate them for no reason. But the key factor is that they love talking about them. So if you build software to disrupt the industry, this has to be in the equation. Your service has to allow people to have a personal connection with the artist. None of the music services now do this. They let you listen to the artist but the connection is already built in from somewhere else. I would say the best right now at doing this is Youtube. But even Youtube could do a lot better job of it. MySpace was kinda on the right path but bad design decisions killed that.
Second, a new service to disrupt the industry would also have to think about the people who make the artist what they are. The stylist, the make up people, the manager, the promoters, the background singers, and etc. An artist is a package. And you would have to allow for that package to be dealt with in a new disruption model. The record labels feed a lot of people and a lot of those people are vital for a record and an artist to be pushed.
I have a lot more that I have to say about this subject but I have already built up a wall of text so I'll cut it here. I'm a little bit passionate about this topic because I have seen for years how the label system screws both artist, DJ's and fans. They need to be brought down and brought down hard but it will not be easy. The one thing that you will find is that most artist have no business sense. They want to be able to create and that's it. We see this line of thinking with some software programmers as well. But programmers usually keep a steady job. Most artist have a short shelf life. It's not until they are broke that they realize how they have been screwed. Before that happens though, they love the industry.","created_at":"2012-01-21T14:24:42Z","created_at_i":1327155882,"objectID":"3493622","parent_id":3492912,"story_id":3492912,"story_title":"Can we kill the music business too?","story_url":"http://audiosearch.blogspot.com/2012/01/can-we-kill-music-business-too.html","updated_at":"2024-09-19T18:18:40Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ajuc"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"In Warsaw Pact countries were assigned specialized roles by USSR. Poland was supposed to make tanks, radars, radio, agricultural machines (including planes), etc. And we had pretty good industry in these domains. But in each case - for example tanks - there were always some parts you couldn't make in one country to prevent uprisings. So for example for tanks we took some crucial parts from Czechoslovakia.
USSR was able to do whatever it wanted of course.
So, to this day we make good military RF-related tech (see WB-Electronics and Radwar for example). But when we tried a space program in 60s ( https://en.wikipedia.org/wiki/Meteor_(rocket) ) - USSR made us stop pretty quickly.
Back to the subject - East Germany and USSR was supposed to do integrated circuits, so our industry wasn't very good in that regard. We made some, but nothing modern. Polish-made computers were mostly using foreign ICs with the exception of a Polish-made clone of 8080 used in a series of computers by Elwro.
Nowadays we have one pretty big IC factory in Poland ( https://en.wikipedia.org/wiki/Wilk_Elektronik ), and a few specialized IC factories for military and industrial automation.
On the other hand Poland is pretty big in software, especially gamedev."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Polish communist era 8 bit computer used in banks, MK-45 (2021) [video]"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.youtube.com/watch?v=CMRAMxtS21A"}},"_tags":["comment","author_ajuc","story_34649461"],"author":"ajuc","comment_text":"In Warsaw Pact countries were assigned specialized roles by USSR. Poland was supposed to make tanks, radars, radio, agricultural machines (including planes), etc. And we had pretty good industry in these domains. But in each case - for example tanks - there were always some parts you couldn't make in one country to prevent uprisings. So for example for tanks we took some crucial parts from Czechoslovakia.
USSR was able to do whatever it wanted of course.
So, to this day we make good military RF-related tech (see WB-Electronics and Radwar for example). But when we tried a space program in 60s ( https://en.wikipedia.org/wiki/Meteor_(rocket) ) - USSR made us stop pretty quickly.
Back to the subject - East Germany and USSR was supposed to do integrated circuits, so our industry wasn't very good in that regard. We made some, but nothing modern. Polish-made computers were mostly using foreign ICs with the exception of a Polish-made clone of 8080 used in a series of computers by Elwro.
Nowadays we have one pretty big IC factory in Poland ( https://en.wikipedia.org/wiki/Wilk_Elektronik ), and a few specialized IC factories for military and industrial automation.
On the other hand Poland is pretty big in software, especially gamedev.","created_at":"2023-02-04T09:25:58Z","created_at_i":1675502758,"objectID":"34653072","parent_id":34649739,"story_id":34649461,"story_title":"Polish communist era 8 bit computer used in banks, MK-45 (2021) [video]","story_url":"https://www.youtube.com/watch?v=CMRAMxtS21A","updated_at":"2024-09-20T13:15:36Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"tptacek"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"Or, \"How not to make money selling software: a succinct illustration of cost-based and market-based (specifically, value-based) pricing in just two comment threads.\"
Why can Github charge FIVE THOUSAND DOLLARS PER YEAR for a local install of Github and succeed, impressively so, despite a small army of nerds pointing out how inexpensive it is to run one's own Git server and Gitorious? Why does 37signals have an office with walls made of 37signalsium (really, seen it, it's fuzzy) and trendy furniture despite selling software that the nerdosphere can clearly duplicate? Why does Yammer even consider publishing a $5/user/month price for software that is among every web geek's top-5-most-likely-personal projects?
Answer: they don't sell gypsum.
Cost based pricing, which works for gypsum sales but not so much for software, suggests that the price of a nice photo should be the price of the gas to get to and from the photo shoot, possibly divided by the number of people interested in buying the photo, plus maybe throw a couple bucks in there and buy yourself something nice, photographer.
Value-based pricing says, \"how much it cost me to create the photo is irrelevant\". YES YES A THOUSAND TIMES YES, say the nerds. THAT'S EXACTLY WHAT I WAS SAYING! But value-based pricing continues: \"no, what matters is how much it would cost you to make that photo and how much benefit it brings\".
How much it would cost you to make that photo: (say) $6000, perhaps divided by the number of different photos you'd take given the same setup (but then also scaled back up to account for the headcount or professional services required to take lots of pictures).
How much benefit? It depends. I could take 10,000 words listing out factors in figuring it out. Most importantly: what are the substitutes to this photo and how much do they cost? For some businesses, clip art of the Eiffel Tower suffices to bring 80-90% of the value. For others (like ad-sales print publications), comparables might also be very expensive.
Now, the extent to which the YOUR business benefit from my photo exceeds MY cost to produce the photo is MY ADVANTAGE IN PRODUCING PHOTOS (the extent to which your cost exceeds my cost is my \"comparative advantage\"\u2020\u2020\u2020, and if it's a positive number possibly suggests that I'm the one who should be doing the photos in any case, since you have better ways to put your money to use). Having an advantage is a good thing. Among other things, it's a key reason why software startups are lucrative, and why we don't all work on line-of-business software for non-software companies.
It's probably true that a photo of a sunset isn't worth $6,000. But, exclusivity aside, the value of the photo also has nothing to do with how much it cost the photographer to take each shot at the margin, and it has nothing to do with the cost to make each marginal sale. What matters is how much it costs the customer to replace that value with a substitute, and in that analysis the $6000 set-up cost, while not determinative, is relevant perspective.
The moralism in these threads is an irrelevant sideshow. Situationally, the nerdosphere oscillates between extremes when trying to compute valuations for stuff with intangible-seeming benefits. Today, the nerdosphere apparently thinks either (a) every photo is a precious snowflake or (b) photo costs should be scaled by the current price of hard disk storage. Yesterday, it was whether it's right for Github to charge per repo. Before that, it's whether it's fair to have markets for spec work like 99designs.
None of that matters. What matters is, is there a market for what you're selling, and will it clear based on the model you use to price stuff on the market. Clearly there is a market for high-quality photography. Clearly it is not a cost-based market like gypsum, or there would not exist sites selling photos with royalties attached, or photos costing hundreds of dollars --- which clearly those sites do exist. So instead of arguing about how much photos should cost --- because, again, they cost what the market says they cost, not what you think it costs to make them yourself --- think instead about how this discussion applies to your own work product. More than you think it does, is my guess.
\u2020\u2020\u2020 (Actually this isn't all what comparative advantage is; comparative advantage says, if there's a market for widgets and a market for photos and you're better at widgets than photos and I'm better at photos than widgets, then I should do widgets and you should do photos, which is a subtly different idea, but the point stands either way.)"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"This photograph is free"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"http://standblog.org/blog/post/2012/01/11/This-photograph-is-free"}},"_tags":["comment","author_tptacek","story_3452908"],"author":"tptacek","children":[3453740,3454094,3454307,3455410,3456054,3456546],"comment_text":"Or, \"How not to make money selling software: a succinct illustration of cost-based and market-based (specifically, value-based) pricing in just two comment threads.\"
Why can Github charge FIVE THOUSAND DOLLARS PER YEAR for a local install of Github and succeed, impressively so, despite a small army of nerds pointing out how inexpensive it is to run one's own Git server and Gitorious? Why does 37signals have an office with walls made of 37signalsium (really, seen it, it's fuzzy) and trendy furniture despite selling software that the nerdosphere can clearly duplicate? Why does Yammer even consider publishing a $5/user/month price for software that is among every web geek's top-5-most-likely-personal projects?
Answer: they don't sell gypsum.
Cost based pricing, which works for gypsum sales but not so much for software, suggests that the price of a nice photo should be the price of the gas to get to and from the photo shoot, possibly divided by the number of people interested in buying the photo, plus maybe throw a couple bucks in there and buy yourself something nice, photographer.
Value-based pricing says, \"how much it cost me to create the photo is irrelevant\". YES YES A THOUSAND TIMES YES, say the nerds. THAT'S EXACTLY WHAT I WAS SAYING! But value-based pricing continues: \"no, what matters is how much it would cost you to make that photo and how much benefit it brings\".
How much it would cost you to make that photo: (say) $6000, perhaps divided by the number of different photos you'd take given the same setup (but then also scaled back up to account for the headcount or professional services required to take lots of pictures).
How much benefit? It depends. I could take 10,000 words listing out factors in figuring it out. Most importantly: what are the substitutes to this photo and how much do they cost? For some businesses, clip art of the Eiffel Tower suffices to bring 80-90% of the value. For others (like ad-sales print publications), comparables might also be very expensive.
Now, the extent to which the YOUR business benefit from my photo exceeds MY cost to produce the photo is MY ADVANTAGE IN PRODUCING PHOTOS (the extent to which your cost exceeds my cost is my \"comparative advantage\"\u2020\u2020\u2020, and if it's a positive number possibly suggests that I'm the one who should be doing the photos in any case, since you have better ways to put your money to use). Having an advantage is a good thing. Among other things, it's a key reason why software startups are lucrative, and why we don't all work on line-of-business software for non-software companies.
It's probably true that a photo of a sunset isn't worth $6,000. But, exclusivity aside, the value of the photo also has nothing to do with how much it cost the photographer to take each shot at the margin, and it has nothing to do with the cost to make each marginal sale. What matters is how much it costs the customer to replace that value with a substitute, and in that analysis the $6000 set-up cost, while not determinative, is relevant perspective.
The moralism in these threads is an irrelevant sideshow. Situationally, the nerdosphere oscillates between extremes when trying to compute valuations for stuff with intangible-seeming benefits. Today, the nerdosphere apparently thinks either (a) every photo is a precious snowflake or (b) photo costs should be scaled by the current price of hard disk storage. Yesterday, it was whether it's right for Github to charge per repo. Before that, it's whether it's fair to have markets for spec work like 99designs.
None of that matters. What matters is, is there a market for what you're selling, and will it clear based on the model you use to price stuff on the market. Clearly there is a market for high-quality photography. Clearly it is not a cost-based market like gypsum, or there would not exist sites selling photos with royalties attached, or photos costing hundreds of dollars --- which clearly those sites do exist. So instead of arguing about how much photos should cost --- because, again, they cost what the market says they cost, not what you think it costs to make them yourself --- think instead about how this discussion applies to your own work product. More than you think it does, is my guess.
\u2020\u2020\u2020 (Actually this isn't all what comparative advantage is; comparative advantage says, if there's a market for widgets and a market for photos and you're better at widgets than photos and I'm better at photos than widgets, then I should do widgets and you should do photos, which is a subtly different idea, but the point stands either way.)","created_at":"2012-01-11T20:48:12Z","created_at_i":1326314892,"objectID":"3453606","parent_id":3452908,"points":null,"story_id":3452908,"story_title":"This photograph is free","story_url":"http://standblog.org/blog/post/2012/01/11/This-photograph-is-free","updated_at":"2023-09-06T20:57:15Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"paulmd"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["8090","software","factory"],"value":"This is a legacy thing, NVIDIA now does binning themselves. It started with Turing and the "A" chips, now it's evolved with Ampere. Partners will buy chips that are already pre-binned into "bad/good/exceptional" tiers (bin 0, bin 1, bin 2). As a result, partners generally no longer apply any special binning to most models of cards, since that mostly covers what they need to know.
And for the Ampere generation, they hardly built any of the lower-tier card models, and those chips have to go somewhere. So they end up using low-tier chips in FTW3/Strix/Gaming X and other high-tier products too. So buying a high-tier model is not any particular guarantee of chip quality anymore either - you can buy a 3090 FTW3 and get a bin-0 chip in it. The only exception is true halo products like the Kingpin/HOF/etc targeted at the XOC market which will definitely get Bin-2 chips and maybe additional binning on top of that. But partners shipped 27 quadrillion FTW3 and Strix and other "flagship" cards and couldn't put good chips in every one of them.
For the mass-market (non-XOC market), everything goes onto the same PCB and you test the finished product and the better ones get the "OC" model sticker. That's it, that's entire binning process for most partners/most of the products on the market. You can have a FTW3 with a bin-0 chip in it. Why do anything more, if you know it's going to sell in 0.1 milliseconds anyway?
The partners that still do binning, typically on limited segments like the Kingpin/HOF/etc, no longer actually test the chip performance, they test the electrical characteristics and then apply statistical analysis to figure out which of the inter-pin voltages/etc are statistically predictive of chip quality. But there is less and less of this because it's expensive and time-consuming, and the differences don't tend to produce significant gains in finished product performance as much as they used to, due to GPU Boost algorithms getting better, etc.
Much like Silicon Lottery exiting the CPU binning business, it's become a bit of a time-waster and partners no longer really bother with it like they used to.
https://www.igorslab.de/en/chip-is-not-equal-to-chip-first-i...
But again, that gets back to the problem with AIB partners: "so what, exactly, would you say you do here?". They made sense in a world where they were allowed some freedom to experiment and service niche product segments/etc. With super tight control, the only thing they are allowed to do anymore is pump up the TDP to try and squeeze out a win, and bigger coolers. Can't do certain kinds of smaller coolers that might compete with the server business either. They just crank up the TDP and slap a bigger cooler on it and crank the price up 20%, it's not really all that significant a benefit to consumers anymore.
They no longer bin. They don't modify the silicon or software in any way. NVIDIA even provides the memory as part of the bundle with the GPU sale. A lot of them use NVIDIA's reference PCB design, and the only common change to that design is custom VRM, which is pointless in a world where overclocking is dead and everyone is undervolting. Cooler design is all pretty functionally-interchangeable across brands, you need something, but a Gaming X is not really different from a FTW3 in any meaningful way. Everything meaningful is done by NVIDIA now.
They are supposed to "buffer inventory" but in practice every time there's a glut they come back to NVIDIA with their hands out looking for refunds for their leftovers and kickbacks on the inventory they keep. So heads they win, tails NVIDIA loses. And it's not like their prices are awesome the rest of the time either.
They no longer even act as a price anchor... 10-series launch, only EVGA and Zotac put out cards anywhere near the "official" MSRP (in hideously limited quantities) and they were still $20-30 over. Everything else was actually higher than the FE cards, in the $725-800 range. Same for the 30-series where NVIDIA cards were the only remotely affordable ones. So these days it's NVIDIA price-anchoring the partners, partners want a 20% higher street price over the NVIDIA design for what seemingly amounts to a higher factory TDP configuration and a bigger cooler. (Which is of course tied to the BOM cost which NVIDIA controls, but... the price structure kinda doesn't matter to the consumer.)
Most of them don't even build the cards themselves... EVGA for example contracts out their assembly to third-party factories in Taiwan, they don't own the factory. So they basically just take the spec from the customer, and give it to the engineers. Well, at least their secretary does that I guess.
So what, exactly, do partners do around here anymore? I realize that a lot of this is the result of product strategy by NVIDIA, some other parts are just the direction that tech has gone over the years (automatic boosting, undervolting, etc), but, regardless, consumers don't really benefit from the things that AIBs are able to offer nowadays. Partners provide almost no real value-add anymore. Functionally they are middlemen, and middlemen get squeezed to zero by the market.
Ironically, in true Office Space fashion, probably the strongest argument is that they deal with the fucking customers so NVIDIA doesn't have to, EVGA could still distinguish itself on warranty/etc. But that's really the only meaningful distinction between AIB partners anymore and it was really only a couple of them that had worthwhile service. With EVGA gone, that argument starts to fall apart too."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"EVGA terminates Nvidia partnership [video]"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.youtube.com/watch?v=cV9QES-FUAM"}},"_tags":["comment","author_paulmd","story_32870677"],"author":"paulmd","comment_text":"This is a legacy thing, NVIDIA now does binning themselves. It started with Turing and the "A" chips, now it's evolved with Ampere. Partners will buy chips that are already pre-binned into "bad/good/exceptional" tiers (bin 0, bin 1, bin 2). As a result, partners generally no longer apply any special binning to most models of cards, since that mostly covers what they need to know.
And for the Ampere generation, they hardly built any of the lower-tier card models, and those chips have to go somewhere. So they end up using low-tier chips in FTW3/Strix/Gaming X and other high-tier products too. So buying a high-tier model is not any particular guarantee of chip quality anymore either - you can buy a 3090 FTW3 and get a bin-0 chip in it. The only exception is true halo products like the Kingpin/HOF/etc targeted at the XOC market which will definitely get Bin-2 chips and maybe additional binning on top of that. But partners shipped 27 quadrillion FTW3 and Strix and other "flagship" cards and couldn't put good chips in every one of them.
For the mass-market (non-XOC market), everything goes onto the same PCB and you test the finished product and the better ones get the "OC" model sticker. That's it, that's entire binning process for most partners/most of the products on the market. You can have a FTW3 with a bin-0 chip in it. Why do anything more, if you know it's going to sell in 0.1 milliseconds anyway?
The partners that still do binning, typically on limited segments like the Kingpin/HOF/etc, no longer actually test the chip performance, they test the electrical characteristics and then apply statistical analysis to figure out which of the inter-pin voltages/etc are statistically predictive of chip quality. But there is less and less of this because it's expensive and time-consuming, and the differences don't tend to produce significant gains in finished product performance as much as they used to, due to GPU Boost algorithms getting better, etc.
Much like Silicon Lottery exiting the CPU binning business, it's become a bit of a time-waster and partners no longer really bother with it like they used to.
https://www.igorslab.de/en/chip-is-not-equal-to-chip-first-i...
But again, that gets back to the problem with AIB partners: "so what, exactly, would you say you do here?". They made sense in a world where they were allowed some freedom to experiment and service niche product segments/etc. With super tight control, the only thing they are allowed to do anymore is pump up the TDP to try and squeeze out a win, and bigger coolers. Can't do certain kinds of smaller coolers that might compete with the server business either. They just crank up the TDP and slap a bigger cooler on it and crank the price up 20%, it's not really all that significant a benefit to consumers anymore.
They no longer bin. They don't modify the silicon or software in any way. NVIDIA even provides the memory as part of the bundle with the GPU sale. A lot of them use NVIDIA's reference PCB design, and the only common change to that design is custom VRM, which is pointless in a world where overclocking is dead and everyone is undervolting. Cooler design is all pretty functionally-interchangeable across brands, you need something, but a Gaming X is not really different from a FTW3 in any meaningful way. Everything meaningful is done by NVIDIA now.
They are supposed to "buffer inventory" but in practice every time there's a glut they come back to NVIDIA with their hands out looking for refunds for their leftovers and kickbacks on the inventory they keep. So heads they win, tails NVIDIA loses. And it's not like their prices are awesome the rest of the time either.
They no longer even act as a price anchor... 10-series launch, only EVGA and Zotac put out cards anywhere near the "official" MSRP (in hideously limited quantities) and they were still $20-30 over. Everything else was actually higher than the FE cards, in the $725-800 range. Same for the 30-series where NVIDIA cards were the only remotely affordable ones. So these days it's NVIDIA price-anchoring the partners, partners want a 20% higher street price over the NVIDIA design for what seemingly amounts to a higher factory TDP configuration and a bigger cooler. (Which is of course tied to the BOM cost which NVIDIA controls, but... the price structure kinda doesn't matter to the consumer.)
Most of them don't even build the cards themselves... EVGA for example contracts out their assembly to third-party factories in Taiwan, they don't own the factory. So they basically just take the spec from the customer, and give it to the engineers. Well, at least their secretary does that I guess.
So what, exactly, do partners do around here anymore? I realize that a lot of this is the result of product strategy by NVIDIA, some other parts are just the direction that tech has gone over the years (automatic boosting, undervolting, etc), but, regardless, consumers don't really benefit from the things that AIBs are able to offer nowadays. Partners provide almost no real value-add anymore. Functionally they are middlemen, and middlemen get squeezed to zero by the market.
Ironically, in true Office Space fashion, probably the strongest argument is that they deal with the fucking customers so NVIDIA doesn't have to, EVGA could still distinguish itself on warranty/etc. But that's really the only meaningful distinction between AIB partners anymore and it was really only a couple of them that had worthwhile service. With EVGA gone, that argument starts to fall apart too.","created_at":"2022-09-17T18:29:24Z","created_at_i":1663439364,"objectID":"32879960","parent_id":32873996,"story_id":32870677,"story_title":"EVGA terminates Nvidia partnership [video]","story_url":"https://www.youtube.com/watch?v=cV9QES-FUAM","updated_at":"2024-09-20T12:05:52Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"dublin"},"comment_text":{"matchLevel":"none","matchedWords":[],"value":"Good times! Capsule history of the computer world:
Before I ever had a computer, I devoured Popular Electronics and the early computer magazines (I still have a copy of "Interface Age" that includes a "Floppy ROM" - one of those vinyl record disks that used to be bound in magazines that you play on a record player - it contained a minimal BASIC ROM encoded in the "Kansas City format" that was the most common way to save data on cassette tapes, since floppy disks (8" at the time for the small ones!) were still heinously expensive - and what hacker needed to store hundreds of KB when 64K RAM was HUGE? (Only big companies had exotic stuff like megabyte hard disks.)
These magazines featured articles about early computers from long-gone companies like IMSAI (8080), SWTPC (6800), Compucolor, Kaypro, and RCA. (Yes, they had a moderately popular early hobbyist computer board that used their own microprocessor...) It took a while for the market momentum to sort-of settle (maybe 40%) on the S-100 bus, Digital Research's CP/M operating system, and the Z-80, but early on, there were literally hundreds of different kinds of PCs, OSes, and early apps, from spreadsheets (VisiCalc and MultiPlan were the two biggest) and word Processors (WordStar was most popular) to, of course, games! CP/M was the common glue that let apps be consistent across such widely varying hardware.
I learned BASIC programming as a kid on TRS-80 model I in the first computer lab at my school in 1978. The first computer I owned was an early Timex/Sinclair ZX-81, but the first "real" computer I had wasn't until I was a senior in college, when I bought a liquidation special Commodore SX-64 - the "portable" version with a cute as buttons 5" color CRT screen and a 1541 floppy disk built in! I wrote all my senior papers and reports with it, as well as my first several resumes, using one of the most amazing hacks ever written: The amazing GEOS from Berkeley Softworks. This incredible program replicated almost all of the capabilities of the MUCH more expensive state-of-the-art Macintosh (with its far more powerful 68000 CPU) on the C-64's 64KB RAM and already pretty long in the tooth 6502 CPU. It wasn't speedy, but was fast enough to be usable, and included a word processor, calculator, spreadsheet, bitmap drawing/painting program, and more. GEOS was later ported to the PC and was used by a few companies I know of that wanted a GUI on PCs prior to Microsoft coming out with Windows in the late 80s. (I also worked at a sizeable aerospace company in those days that standardized on Digital Research's GEM OS and app suite. We did a fair amount of the automated factory design (except CAD, of course), planning, and documentation for the B-2 airframe in the GEM apps. GEM wasn't for most home users: it required an exotic VGA (640x480) graphics adapter! Interestingly, we though HDTV monitors were right around the corner then, but they wouldn't arrive for another 15 years. Our displays and interfaces still suck, compared to what they should be - think desk-sized touch/pen Surface Studio.)
By the late 80s, the PC was dominant, driven first by the 8086 (with 8087 math coprocessor if you were doing badass stuff) and later by the new and powerful Intel '286, '386, and '486 CPUs. IBM PC compatibility was becoming required, largely driven by a new upstart called Compaq making 30 lb.(!) suitcase computers - They sold like hotcakes. Very early PCs, even IBM's, gave you the choice of PC-DOS (IBM's version of what would later be MS-DOS) or CP/M-86 (the preference of us "serious and knowledgeable" computer guys - when spending someone else's money: it cost several times more than DOS!) CP/M-86 felt like a cross between CP/M and early DEC OSes - much more powerful than the clunky DOS. To do more than one thing, you had to load programs called TSRs (Terminate and Stay Resident), the most popular of which was Borland's Sidekick, which was probably the most pirated software on the planet for a while. Running multiple TSRs was possible, but they often got along like strange bulldogs in a bag... The PC software world changed forever when Lotus introduced 1-2-3, its lightning fast spreadsheet written in C, the language of serious Unix computers, which allowed nearly the speed of assembly (which was behind most commercial apps), but could be written by mere mortals who happened to be geniuses.
Macs were cool, but really expensive (like literally, the price of a new car) and more like toys than tools, unless you were doing desktop publishing, where the LaserWriter changed everything - I honestly don't think Apple would have succeeded with the Mac at all without the LaserWriter - it was that important.
Things started to get standard by the early 90s, when Windows 3 finally brought GUIs to the masses at a reasonable cost, and allowed Microsoft to extend its OS dominance into apps, where they had been a player, but not dominant. Things stayed that way until I saw a VERY interesting demo at Interop '93 (IIRC) of O'Reilly's GNN - Global Network Navigator - the first news and info website anyone had ever seen designed to work with the stunning new NCSA Mosaic browser that blew everyone away with its support for inline graphics (early web browsers required you to click on links and download the image to see it in a separate window.) It took a few more years, but that idea was the one that created the modern world, though we had to go through a few awkward years of Trumpet Winsock to add TCP/IP to Windows before Windows 97 showed up with integrated IP support.
I've skipped tons, of course, but the rest has pretty much just evolved from there into what you see today..."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Who where lived through the beginning of the personal computer era?"}},"_tags":["comment","author_dublin","story_31673567"],"author":"dublin","comment_text":"Good times! Capsule history of the computer world:
Before I ever had a computer, I devoured Popular Electronics and the early computer magazines (I still have a copy of "Interface Age" that includes a "Floppy ROM" - one of those vinyl record disks that used to be bound in magazines that you play on a record player - it contained a minimal BASIC ROM encoded in the "Kansas City format" that was the most common way to save data on cassette tapes, since floppy disks (8" at the time for the small ones!) were still heinously expensive - and what hacker needed to store hundreds of KB when 64K RAM was HUGE? (Only big companies had exotic stuff like megabyte hard disks.)
These magazines featured articles about early computers from long-gone companies like IMSAI (8080), SWTPC (6800), Compucolor, Kaypro, and RCA. (Yes, they had a moderately popular early hobbyist computer board that used their own microprocessor...) It took a while for the market momentum to sort-of settle (maybe 40%) on the S-100 bus, Digital Research's CP/M operating system, and the Z-80, but early on, there were literally hundreds of different kinds of PCs, OSes, and early apps, from spreadsheets (VisiCalc and MultiPlan were the two biggest) and word Processors (WordStar was most popular) to, of course, games! CP/M was the common glue that let apps be consistent across such widely varying hardware.
I learned BASIC programming as a kid on TRS-80 model I in the first computer lab at my school in 1978. The first computer I owned was an early Timex/Sinclair ZX-81, but the first "real" computer I had wasn't until I was a senior in college, when I bought a liquidation special Commodore SX-64 - the "portable" version with a cute as buttons 5" color CRT screen and a 1541 floppy disk built in! I wrote all my senior papers and reports with it, as well as my first several resumes, using one of the most amazing hacks ever written: The amazing GEOS from Berkeley Softworks. This incredible program replicated almost all of the capabilities of the MUCH more expensive state-of-the-art Macintosh (with its far more powerful 68000 CPU) on the C-64's 64KB RAM and already pretty long in the tooth 6502 CPU. It wasn't speedy, but was fast enough to be usable, and included a word processor, calculator, spreadsheet, bitmap drawing/painting program, and more. GEOS was later ported to the PC and was used by a few companies I know of that wanted a GUI on PCs prior to Microsoft coming out with Windows in the late 80s. (I also worked at a sizeable aerospace company in those days that standardized on Digital Research's GEM OS and app suite. We did a fair amount of the automated factory design (except CAD, of course), planning, and documentation for the B-2 airframe in the GEM apps. GEM wasn't for most home users: it required an exotic VGA (640x480) graphics adapter! Interestingly, we though HDTV monitors were right around the corner then, but they wouldn't arrive for another 15 years. Our displays and interfaces still suck, compared to what they should be - think desk-sized touch/pen Surface Studio.)
By the late 80s, the PC was dominant, driven first by the 8086 (with 8087 math coprocessor if you were doing badass stuff) and later by the new and powerful Intel '286, '386, and '486 CPUs. IBM PC compatibility was becoming required, largely driven by a new upstart called Compaq making 30 lb.(!) suitcase computers - They sold like hotcakes. Very early PCs, even IBM's, gave you the choice of PC-DOS (IBM's version of what would later be MS-DOS) or CP/M-86 (the preference of us "serious and knowledgeable" computer guys - when spending someone else's money: it cost several times more than DOS!) CP/M-86 felt like a cross between CP/M and early DEC OSes - much more powerful than the clunky DOS. To do more than one thing, you had to load programs called TSRs (Terminate and Stay Resident), the most popular of which was Borland's Sidekick, which was probably the most pirated software on the planet for a while. Running multiple TSRs was possible, but they often got along like strange bulldogs in a bag... The PC software world changed forever when Lotus introduced 1-2-3, its lightning fast spreadsheet written in C, the language of serious Unix computers, which allowed nearly the speed of assembly (which was behind most commercial apps), but could be written by mere mortals who happened to be geniuses.
Macs were cool, but really expensive (like literally, the price of a new car) and more like toys than tools, unless you were doing desktop publishing, where the LaserWriter changed everything - I honestly don't think Apple would have succeeded with the Mac at all without the LaserWriter - it was that important.
Things started to get standard by the early 90s, when Windows 3 finally brought GUIs to the masses at a reasonable cost, and allowed Microsoft to extend its OS dominance into apps, where they had been a player, but not dominant. Things stayed that way until I saw a VERY interesting demo at Interop '93 (IIRC) of O'Reilly's GNN - Global Network Navigator - the first news and info website anyone had ever seen designed to work with the stunning new NCSA Mosaic browser that blew everyone away with its support for inline graphics (early web browsers required you to click on links and download the image to see it in a separate window.) It took a few more years, but that idea was the one that created the modern world, though we had to go through a few awkward years of Trumpet Winsock to add TCP/IP to Windows before Windows 97 showed up with integrated IP support.
I've skipped tons, of course, but the rest has pretty much just evolved from there into what you see today...","created_at":"2022-06-08T23:35:21Z","created_at_i":1654731321,"objectID":"31675855","parent_id":31673567,"story_id":31673567,"story_title":"Ask HN: Who where lived through the beginning of the personal computer era?","updated_at":"2024-09-20T11:21:10Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"xoa"},"comment_text":{"matchLevel":"none","matchedWords":[],"value":">Why is it in general required, to design a new, more heavy rocket - and not "just scale" up an existing design?
I'm not a rocket designer, just someone who has really enjoyed following it for a long time. Someone in the industry could give more specifics, but the basics are straight forward. Everything about a specific rocket is tightly coupled, from material factors to engines and everything that goes into that. To take an obvious one, how do you drive the bigger rocket? If you choose bigger engines, then you need entire new engines (and there are scaling issues there too). If you choose to use more smaller engines, you need brand new plumbing to feed all those engines (a very complex thing itself, last I checked each Raptor engine consumers nearly 800 kilograms of fuel PER SECOND), you need to deal with harmonics and control of more engines, different fluid mechanics, etc. Even merely construction changes with scale. For example scaling up carbon fiber to huge sizes is incredibly hard and expensive, one of a number of reasons SpaceX dropped it in favor of steel instead. Rocket volume obviously scales with the cube compared to surface area with the square of size, material properties change with thickness of various walls, and on and on. All of this also impacts aerodynamics, center of gravity and a host of other things.
So yeah! Knowledge and expertise from existing rocket designs is absolutely very useful in designing a new one, Starship directly comes out of a ton of things SpaceX learned from F9, which itself took important lessons from F1 (and "F9" is a much bigger range much faster then typical rockets, from version 1.0 to Full Thrust Block 5, the current one, they changed more and gained more power then entire new rockets sometimes have). A much bigger Starship v2 may take much more from v1 then typical because they now do have something with more scaling potential (steel construction in particular). And a lot of the software no doubt will be directly evolved, Starship landing of course used F9 landing software and hardware lessons. But still, cannot just "scale up", even without things like completely changing fuel.
Edit: In reply, jmyeet points out that I glossed over the manufacturing bit too much as well. There is a huge amount of expertise and institutional knowledge in going from a general design of any big complex machine to actually manufacturing it in a reliable fashion. People and the machines that make the machines and all the processes that go into that is a whole 'nother ball of wax. Can find plenty of stories on HN and elsewhere about merely duplicating "the same thing", like trying to manufacture an aircraft in a different factory. Or the level of work in "simply" copying a fab and the lengths that are gone to in order to try to make sure everything is exactly the same."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Starship is threatening NASA\u2019s moon contractors"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.politico.com/news/2022/02/12/elon-musk-space-freaking-out-competitors-00008441"}},"_tags":["comment","author_xoa","story_30342446"],"author":"xoa","children":[30347601,30355690],"comment_text":">Why is it in general required, to design a new, more heavy rocket - and not "just scale" up an existing design?
I'm not a rocket designer, just someone who has really enjoyed following it for a long time. Someone in the industry could give more specifics, but the basics are straight forward. Everything about a specific rocket is tightly coupled, from material factors to engines and everything that goes into that. To take an obvious one, how do you drive the bigger rocket? If you choose bigger engines, then you need entire new engines (and there are scaling issues there too). If you choose to use more smaller engines, you need brand new plumbing to feed all those engines (a very complex thing itself, last I checked each Raptor engine consumers nearly 800 kilograms of fuel PER SECOND), you need to deal with harmonics and control of more engines, different fluid mechanics, etc. Even merely construction changes with scale. For example scaling up carbon fiber to huge sizes is incredibly hard and expensive, one of a number of reasons SpaceX dropped it in favor of steel instead. Rocket volume obviously scales with the cube compared to surface area with the square of size, material properties change with thickness of various walls, and on and on. All of this also impacts aerodynamics, center of gravity and a host of other things.
So yeah! Knowledge and expertise from existing rocket designs is absolutely very useful in designing a new one, Starship directly comes out of a ton of things SpaceX learned from F9, which itself took important lessons from F1 (and "F9" is a much bigger range much faster then typical rockets, from version 1.0 to Full Thrust Block 5, the current one, they changed more and gained more power then entire new rockets sometimes have). A much bigger Starship v2 may take much more from v1 then typical because they now do have something with more scaling potential (steel construction in particular). And a lot of the software no doubt will be directly evolved, Starship landing of course used F9 landing software and hardware lessons. But still, cannot just "scale up", even without things like completely changing fuel.
Edit: In reply, jmyeet points out that I glossed over the manufacturing bit too much as well. There is a huge amount of expertise and institutional knowledge in going from a general design of any big complex machine to actually manufacturing it in a reliable fashion. People and the machines that make the machines and all the processes that go into that is a whole 'nother ball of wax. Can find plenty of stories on HN and elsewhere about merely duplicating "the same thing", like trying to manufacture an aircraft in a different factory. Or the level of work in "simply" copying a fab and the lengths that are gone to in order to try to make sure everything is exactly the same.","created_at":"2022-02-15T15:07:58Z","created_at_i":1644937678,"objectID":"30347405","parent_id":30347097,"story_id":30342446,"story_title":"Starship is threatening NASA\u2019s moon contractors","story_url":"https://www.politico.com/news/2022/02/12/elon-musk-space-freaking-out-competitors-00008441","updated_at":"2024-09-20T10:31:05Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"RNCTX"},"comment_text":{"matchLevel":"none","matchedWords":[],"value":"> This misses a big point: highly skilled craftsmanship will never again be affordable on an architectural scale. So we'll have it again when it can all be done by machine.
That's not really the case. As with the internet and software, architecture was hijacked by building products corporations and tradespeople were duped into playing along, just like the employees of social media corporations one can see on this site who try to justify the evil activities of the companies they work for.
Background: when I quit software in the latter half of the 2000s I bought a house built in 1908 in the central US. It needed lots of things, so I basically lived in half of it and worked on the other half for about 10 years.
What I learned in the process of learning how to do all of that stuff by trial and error (including building custom windows and doors... wooden ones to replace the 'new' ones from previous owners that I threw away) is that prior to the 1980s there were lots of small machines built for the market of a small custom shop that might build things for local clientele. They have all been replaced by worse (in every respect) multi-function machines that only make boring things badly.
As a baseline, the most notorious of big box stores [1] says that new plastic windows cost 800 - 1000 dollars. That's highway robbery. There's about 30 board feet of rough sawn lumber in a 4' x 3' wooden double hung window. The material cost (assuming cypress in the southeast or douglas fir in the north/west) is gonna be around 5 dollars per board foot. Glass in smallish bulk quantities lets say 3 dollars per square foot, and hardware about 50 dollars per window, I'll give 10-20 dollars for paint and what not. So we're looking at a material cost of around $250.00 per window if someone wanted wooden windows in a common size.
The problem is not that it's impossible to compete with the plastic window on quality at a slight price markup, the problem is that there's no one to do it because everyone has been turned into a middle man, serving rent seekers. Outside of major cities with enough of a concentration of historic buildings to keep real craftsmen busy, there aren't going to be any such real craftsmen. They're all going to be glorified salesmen with a few hand tools from the local building supply retailer.
It didn't have to be this way. As an example, up until the late 1970s a company named Powermatic built a small single end tender (model # 2A) that is shown here [2] and here [3]. These were marketed to small shops, not giant industrial factories. These machines have appreciated in value on the used market in the past 10 years, because people are still using them! Ten years ago you could buy a serviceable 2A tenoner for $500-1000 dollars, now they're going for $2500 to 3k. I met people running window and door shops doing historic preservation work around Boston that were stockpiling every one they find for sale for parts. Powermatic the company on the other hand was sold and the brand "cashed in" on cheap import imitations of other tool companies' products long ago.
Another example would be the machine called a "sash trimmer" shown here [4] and here [5]. I use the term "machine" very loosely because it's not even motorized, you operate it with your foot like an old sewing machine. This machine could be replicated in a machine shop today using an old milling machine or mortiser as a base for probably $3000 dollars or so in custom parts. It's how you make fancy wooden windows like these [6]. It's virtually impossible to make such a wooden window without that machine, because modern machines can't replicate the required angled mortises. I looked for years and never found one for sale in the southeastern US. Despite the fact that thousands of historic homes in the southeastern US have such windows, that are all over 80 years old now and gradually rotting away.
You can extend this to any other item in the typical single family home in the US. For instance there's no reason for kitchen cabinets to have plywood in them. Plywood is another notorious "profit margin" item by the building products corporations that is basically scrap and glue with a picture of real wood on top of it. Why isn't there a set of jigs like the ones used 80+ years ago for sale by which a local cabinet maker can produce "stick built" cabinets without plywood shells? Why does the local cabinet maker pay obscene prices for fake lumber pressed into sheets, when by the square foot solid yellow pine is cheaper? Because the local cabinet shop sees himself as a building product middle man just like the corporate stores do, there is no other reason.
As if the entirety of American architecture being beholden to corporate producers of junk products isn't bad enough, general contractors are charging an additional 20-30% markup (at least) on top of the supplier's markup, to lead people to think that the junk they've bought is reasonably that expensive.
As the wise machine said in the movie from the cursed decade: "the only way to win, is to not play."
Read a few books, go to architectural salvage stores to look at things in person and see how they were built, go to your local woodworker's meetup group, go watch youtube videos about how to glaze windows and hang doors. You don't have to buy a shop full of antique machines and start with rough lumber if you don't want to, but if you know how to spec and order things from local craftsmen on your own without the aforementioned rent seeking middlemen charging you 20-30% to do it on your behalf, you can have a house just as nice as those from the late 19th and early 20th century, albeit with some quality of lumber compromises due to the lack of old growth supply.
1. https://www.homedepot.com/c/cost_install_replacement_windows
2. https://youtu.be/TT13WmJZyU0
3. https://www.thewoodfurniturefactory.com/tour.htm
4. https://www.facebook.com/seattlehistoricwindow/videos/102924...
5. http://vintagemachinery.org/photoindex/detail.aspx?id=9824
6. https://www.google.com/search?q=diamond+window+sash"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"When is the revolution in architecture coming?"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.currentaffairs.org/2021/04/when-is-the-revolution-in-architecture-coming"}},"_tags":["comment","author_RNCTX","story_29155987"],"author":"RNCTX","children":[29164449],"comment_text":"> This misses a big point: highly skilled craftsmanship will never again be affordable on an architectural scale. So we'll have it again when it can all be done by machine.
That's not really the case. As with the internet and software, architecture was hijacked by building products corporations and tradespeople were duped into playing along, just like the employees of social media corporations one can see on this site who try to justify the evil activities of the companies they work for.
Background: when I quit software in the latter half of the 2000s I bought a house built in 1908 in the central US. It needed lots of things, so I basically lived in half of it and worked on the other half for about 10 years.
What I learned in the process of learning how to do all of that stuff by trial and error (including building custom windows and doors... wooden ones to replace the 'new' ones from previous owners that I threw away) is that prior to the 1980s there were lots of small machines built for the market of a small custom shop that might build things for local clientele. They have all been replaced by worse (in every respect) multi-function machines that only make boring things badly.
As a baseline, the most notorious of big box stores [1] says that new plastic windows cost 800 - 1000 dollars. That's highway robbery. There's about 30 board feet of rough sawn lumber in a 4' x 3' wooden double hung window. The material cost (assuming cypress in the southeast or douglas fir in the north/west) is gonna be around 5 dollars per board foot. Glass in smallish bulk quantities lets say 3 dollars per square foot, and hardware about 50 dollars per window, I'll give 10-20 dollars for paint and what not. So we're looking at a material cost of around $250.00 per window if someone wanted wooden windows in a common size.
The problem is not that it's impossible to compete with the plastic window on quality at a slight price markup, the problem is that there's no one to do it because everyone has been turned into a middle man, serving rent seekers. Outside of major cities with enough of a concentration of historic buildings to keep real craftsmen busy, there aren't going to be any such real craftsmen. They're all going to be glorified salesmen with a few hand tools from the local building supply retailer.
It didn't have to be this way. As an example, up until the late 1970s a company named Powermatic built a small single end tender (model # 2A) that is shown here [2] and here [3]. These were marketed to small shops, not giant industrial factories. These machines have appreciated in value on the used market in the past 10 years, because people are still using them! Ten years ago you could buy a serviceable 2A tenoner for $500-1000 dollars, now they're going for $2500 to 3k. I met people running window and door shops doing historic preservation work around Boston that were stockpiling every one they find for sale for parts. Powermatic the company on the other hand was sold and the brand "cashed in" on cheap import imitations of other tool companies' products long ago.
Another example would be the machine called a "sash trimmer" shown here [4] and here [5]. I use the term "machine" very loosely because it's not even motorized, you operate it with your foot like an old sewing machine. This machine could be replicated in a machine shop today using an old milling machine or mortiser as a base for probably $3000 dollars or so in custom parts. It's how you make fancy wooden windows like these [6]. It's virtually impossible to make such a wooden window without that machine, because modern machines can't replicate the required angled mortises. I looked for years and never found one for sale in the southeastern US. Despite the fact that thousands of historic homes in the southeastern US have such windows, that are all over 80 years old now and gradually rotting away.
You can extend this to any other item in the typical single family home in the US. For instance there's no reason for kitchen cabinets to have plywood in them. Plywood is another notorious "profit margin" item by the building products corporations that is basically scrap and glue with a picture of real wood on top of it. Why isn't there a set of jigs like the ones used 80+ years ago for sale by which a local cabinet maker can produce "stick built" cabinets without plywood shells? Why does the local cabinet maker pay obscene prices for fake lumber pressed into sheets, when by the square foot solid yellow pine is cheaper? Because the local cabinet shop sees himself as a building product middle man just like the corporate stores do, there is no other reason.
As if the entirety of American architecture being beholden to corporate producers of junk products isn't bad enough, general contractors are charging an additional 20-30% markup (at least) on top of the supplier's markup, to lead people to think that the junk they've bought is reasonably that expensive.
As the wise machine said in the movie from the cursed decade: "the only way to win, is to not play."
Read a few books, go to architectural salvage stores to look at things in person and see how they were built, go to your local woodworker's meetup group, go watch youtube videos about how to glaze windows and hang doors. You don't have to buy a shop full of antique machines and start with rough lumber if you don't want to, but if you know how to spec and order things from local craftsmen on your own without the aforementioned rent seeking middlemen charging you 20-30% to do it on your behalf, you can have a house just as nice as those from the late 19th and early 20th century, albeit with some quality of lumber compromises due to the lack of old growth supply.
1. https://www.homedepot.com/c/cost_install_replacement_windows
2. https://youtu.be/TT13WmJZyU0
3. https://www.thewoodfurniturefactory.com/tour.htm
4. https://www.facebook.com/seattlehistoricwindow/videos/102924...
5. http://vintagemachinery.org/photoindex/detail.aspx?id=9824
6. https://www.google.com/search?q=diamond+window+sash","created_at":"2021-11-09T07:45:53Z","created_at_i":1636443953,"objectID":"29158888","parent_id":29156272,"story_id":29155987,"story_title":"When is the revolution in architecture coming?","story_url":"https://www.currentaffairs.org/2021/04/when-is-the-revolution-in-architecture-coming","updated_at":"2024-09-20T09:48:54Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"tmountain"},"comment_text":{"matchLevel":"none","matchedWords":[],"value":"Ah yes, I'm well acquainted with the simulator. Unfortunately, the reboot time was on par with that of a physical device which lead me to the drastic measure of snapshotting a running simulator inside of virtual box since I could boot virtual box in 20% of the time it took to start a fresh simulator.
The simulator is also throttled to a paltry 10k/sec when performing network access (why?!), and the app I was working on was for music streaming which forced me to do most of my testing on physical devices.
Regarding the connection factory, I was aware of it; however, given the market share of older devices, it makes sense to support them. If you're going to do so, you're going to have to roll your own connection code since older devices can't use newer OS versions. I'm not talking about ancient devices either. The Curve 8900 shipped with the 4.6 which doesn't support the connection factory.
In our case, it didn't make sense to have two builds (legacy and modern) since we'd have to support legacy anyway. I suspect this will improve as the older devices disappear by way of attrition, but it made things difficult at the time.
Regarding the people willing to spend money on software, we never applied to get into the marketplace; however, all of RIM's competitors have similar offerings, so I see no advantage here."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Open letter to BlackBerry bosses: Senior RIM exec tells all as company crumbles"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"http://www.bgr.com/2011/06/30/open-letter-to-blackberry-bosses-senior-rim-exec-tells-all-as-company-crumbles-around-him/"}},"_tags":["comment","author_tmountain","story_2714270"],"author":"tmountain","children":[2716671],"comment_text":"Ah yes, I'm well acquainted with the simulator. Unfortunately, the reboot time was on par with that of a physical device which lead me to the drastic measure of snapshotting a running simulator inside of virtual box since I could boot virtual box in 20% of the time it took to start a fresh simulator.
The simulator is also throttled to a paltry 10k/sec when performing network access (why?!), and the app I was working on was for music streaming which forced me to do most of my testing on physical devices.
Regarding the connection factory, I was aware of it; however, given the market share of older devices, it makes sense to support them. If you're going to do so, you're going to have to roll your own connection code since older devices can't use newer OS versions. I'm not talking about ancient devices either. The Curve 8900 shipped with the 4.6 which doesn't support the connection factory.
In our case, it didn't make sense to have two builds (legacy and modern) since we'd have to support legacy anyway. I suspect this will improve as the older devices disappear by way of attrition, but it made things difficult at the time.
Regarding the people willing to spend money on software, we never applied to get into the marketplace; however, all of RIM's competitors have similar offerings, so I see no advantage here.","created_at":"2011-06-30T20:37:24Z","created_at_i":1309466244,"objectID":"2715653","parent_id":2715050,"story_id":2714270,"story_title":"Open letter to BlackBerry bosses: Senior RIM exec tells all as company crumbles","story_url":"http://www.bgr.com/2011/06/30/open-letter-to-blackberry-bosses-senior-rim-exec-tells-all-as-company-crumbles-around-him/","updated_at":"2024-09-19T17:47:43Z"}],"hitsPerPage":20,"nbHits":79,"nbPages":4,"page":0,"params":"query=8090+software+factory&advancedSyntax=true&analyticsTags=backend","processingTimeMS":41,"processingTimingsMS":{"_request":{"roundTrip":22},"afterFetch":{"format":{"highlighting":5,"total":5},"merge":{"entries":{"decompress":6,"total":6},"mergeLoop":{"prepareNextHit":15,"total":15},"total":16},"total":16},"fetch":{"query":12,"scanning":12,"total":25},"total":42},"query":"8090 software factory","serverTimeMS":48}