6 ms·

The article keeps referring to "AI" as if it's a coherent, agreed-upon thing. It's not. "AI" is a marketing term that's been applied to whatever happens to loo

The article keeps referring to "AI" as if it's a coherent, agreed-upon thing. It's not. "AI" is a marketing term that's been applied to whatever happens to look impressive at the moment---right now, that's LLMs. There's no settled definition, no unified discipline, and lumping everything from generative text models to actual robotics under the same label just obscures the discussion. If we're talking about LLMs, let's call them LLMs. And let's be real: we're in a bubble. There's zero evidence that LLMs have produced a measurable productivity boom. If anything, they're net negative in many contexts---encouraging shallow engagement, short-circuiting the learning process, and making knowledge workers more dependent on stochastic parrots that frequently hallucinate. We've replaced "look it up" with "ask the oracle," and the oracle can't reliably tell fact from fiction. The hype cycle will keep burning cash until someone admits the emperor's wardrobe is ... speculative at best.

edg5000 1y ago
I encourage you to try and really lean on LLMs (Claude Code in particular), it's hard to argue that there is no productivity boost. Athropy in "low level" skills is certainly real though. But our ancestors knew how to repair a pair of socks. Due to industrialisation, we just buy new ones and have lost the skills to repair them. Most people may forget how to do many things in programming, but may still intimately understand, at a high level, what each line is doing.
- bamboozled 1y ago
  The more you use it and the deeper you go the more you realise its limitations as a useful software engineering tool and more as an autocomplete. Is it useful ? Yes? Am I happy it exists yes ? Is it revolutionary. I doubt it.
  - scotty79 1y ago
    Year ago with enough prompting it could write a function for me that did something conceptually simple but tedious. I tried Windsurf over the last 3 days. It can write complex game mechanics logic in multiple files and refactor code when told to. I can talk to it using language of the domain I just made up and I can talk to it using language that developers use when they talk about the code. It can solve bugs that would take me half an hour to figure out at least. It went from a poor junior to clever junior and half a senior in one year. It's not great at one shot architecting yet, but when told how to architect a thing it can move stuff around to conform with the desired architecture. It creates and modifies SVG's, understands colors and directions. It does all of that blind, without running the code. All it has is compiler (language server really) and linter. You mention it's an autocomplete. I barely used auto-complete at all. I just told it what to do. I touched the code a little bit only because I wanted to, never because I had to. When I wanted a change to the code, 95% of the time I just told it what change I needed and it made it. It basically replaces about 1/3 of software team already enabling (and pushing me towards) a role of (tech lead/architect/qa/product owner) and this fraction is rising. It's not flawless. Sometimes it makes things that don't work on the first try. But you don't always need to revert. It's capable of fixing what it made when told about how the desired behavior differs. And all that without running the program. I anticipate that further development is going to be letting it run and observe what it built so it's gonna move towards QA role. And the other thing might be slowly learning to differentiate between well architected code and badly architected code but that's probably harder as it requires careful preparation (and creation) of training data. While moving towards QA just requires figuring out how to let it run and inspect stuff. Maybe use the debugger.
    - bamboozled 1y ago
      Are you familiar with the garnet hype cycle? https://www.gartner.com/en/research/methodologies/gartner-hype-cycle https://www.gartner.com/en/research/methodologies/gartner-hy... I'm doing everything you're doing and probably more with Claude code (I 'let it run'). I'd say I too have been through a few of the different phases of the gartner hype cycle now. As I said it's a good tool, it's far from being an "auto-pilot".
      - scotty79 1y ago
        I don't care about the hype. If you put me in a cave for last 5 years and today gave me Windsurf and told me nothing beyond that a machine does that, I wouldn't believe you. All of it was literally science fiction 5 years ago. This literally replaces a person that I would have to hire to pursue my ideas.
        bamboozled 1y ago
        Ok I guess that's where we differ, I don't really find it to be science fiction, the more I use it the system, the more I understand that magic of it, what it's going to do, and I've learned that to get good results, you absolutely have to drive it like almost every other tool. I guess if anyone went into a cave and came out for 5 years things would be like science fiction. FPV drones on the battlefield would be another example... As I said earlier, not downplaying it, really just warning you that if you lean to heavy on it without guiding it quite precisely, you will get burned, properly. All good, we have different opinions.
        scotty79 1y ago
        FPV drones on the battlefield were absolutely obvious. The only strange thing about it was, and still is, that powerful armies of the biggest players were "nah, I'll pass" on this obvious application. Wedding photographer figured out that you can hit people with drones. Farmers figured out they can carry and release significant payloads. Yet for some reason, generals seen them as toys and preferred to stick with their bombs and artillery shells costing order of magnitude more. We needed a war and soldiers on the contact line to basically start doing this obvious thing out of necessity and demanding supply from commanders. Computers being able to keep the conversation going about complex topic like development of a computer program were 5 years ago farther away than fusion energy. Perpetually 30 to 100 years away. Sure, for the best results you need to guide it because without it you'll get some program but not a good program. If you don't help it to control the complexity of the code with sensible architecture it will paint itself into a corner. But a junior developer would also do that if you just piled up requirements and never steered his output. I agree with you that it's a tool. But more like a combine harvester, that you still need to drive, for now, than a hammer. Already more like a CNC machine than a power drill. And still getting better. Everyone can have their own opinion on this. But I think next 5 years will change those opinions. For me it is at least as magical as cell phones. When I was a kid the peek of wireless personal communication was walkie-talkie with a very limited range. The idea that you could cram enough electronics, RF stuff and a battery in a small enough package and route the calls globally with vast networks of two way communication stations seemed absolutely ludicrous. I haven't even seen SMD element before I had my first cellphone and I picked apart a lot of electronics. Then it happened. It was the kind of technology that got me completely surprised, like LLMs. I remember not being that shocked because I was young and was expecting new things. I was reading a lot of SF and there instant wireless communication was everywhere. But after 4 decades of basically no progress in the machines that do anything that resembles thinking this innovation caught me off guard completely. I would be less surprised if we already had fusion reactors providing most of our electricity. On a more practical note, do you switch a lot the models that you use? I can't really tell them apart because I haven't worked with them all that much but I heard good things about GPT5 and when I used it it was very good (but others, like Claude or swt-1 were great too).
- laserlight 1y ago
  > it's hard to argue that there is no productivity boost I encourage you to keep up with the literature [0]. [0] https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/ https://metr.org/blog/2025-07-10-early-2025-ai-experienced-o...
jama211 1y ago
Personally I now get my work done in half the time and spend the rest of my newfound time with my family. It’s gonna be hard to convince me that isn’t a fantastic productivity increase.
- eru 1y ago
  That's interesting! Historically (at least for the last century or so) we mostly stopped lowering working hours and instead focused on increased output. Eg people still work eight hours a day, but get paid vastly more in real terms than in the 1950s. Instead of working only one hour a day (or fewer years or whatever) and taking a 1950s compensation package. Of course, this is merely a statistical observation. There are plenty of people who eg decide to retire early on a modest nest egg.
  - actionfromafar 1y ago
    Funny how a 1950s compensation package could pay for a house and family though.
    - eru 1y ago
      1950s houses are no longer legal in most places. Both because in the US new construction is basically outlawed. But also because 1950s style houses and goods in general are so shoddy, you can no longer legally produce or buy them. In many cases, no one would even take them off your for free, even if they were still legal.) On a global scale, the times between perhaps 1940s to 1970s had some of the harshest inequality ever. It's been since about the 1980s that inequality has gone down markedly. Remember how India and China used to be on the verge of famine (or outright in famine). Nowadays obesity is the bigger problem.
  - jama211 1y ago
    Remote work has enabled hours to be lowered again for many people, it’s just no one wants to say the quiet part out loud about that.
    - 1y ago
OtherShrezzing 1y ago
> There's zero evidence that LLMs have produced a measurable productivity boom I’m not so sure on this one. AlphaFold2, while not quite an LLM, was based on Transformer architecture - and its implementation wasn’t a million miles away from a language model - and it massively improved the rate of protein structure prediction. I think in general you’re correct, that we’re in a bubble, but I think it’s too extreme to say the technology is valueless.
- JohnKemeny 1y ago
  From what I've read, AlphaFold2 is a minor improvement over other solvers, and only in some cases. It makes many more mistakes and misleading results than classic solvers. But don't take my word for it, it's way outside of my area of research.
eru 1y ago
> And let's be real: we're in a bubble. There's zero evidence that LLMs have produced a measurable productivity boom. [...] You are right that a general productivity boom is hard to detect so far. But we already have certain industries with more than measurable productivity impact. A very salient example: first level call centre agents. The kind of unfortunate souls who are paid to man a phone and not allowed to deviate from their script, nor authorised to actually solve your problem. The companies renting out these services make heavy use of current AI and that has measurably affected hiring, staffing, prices etc. Even fairly basic AI (like what we have today) is good for this use case. Does it suck: sure. Does it suck more than the status quo with warm bodies without power: not really. And it's cheaper.
tim333 1y ago
"AI" is a marketing term... It's also been a coherent academic discipline since at least Turing's 1950 paper introducing the Turing test and with a history going back well beyond that.

edg5000 1y ago

I encourage you to try and really lean on LLMs (Claude Code in particular), it's hard to argue that there is no productivity boost. Athropy in "low level" skills is certainly real though. But our ancestors knew how to repair a pair of socks. Due to industrialisation, we just buy new ones and have lost the skills to repair them. Most people may forget how to do many things in programming, but may still intimately understand, at a high level, what each line is doing.

bamboozled 1y ago
The more you use it and the deeper you go the more you realise its limitations as a useful software engineering tool and more as an autocomplete. Is it useful ? Yes? Am I happy it exists yes ? Is it revolutionary. I doubt it.
- scotty79 1y ago
  Year ago with enough prompting it could write a function for me that did something conceptually simple but tedious. I tried Windsurf over the last 3 days. It can write complex game mechanics logic in multiple files and refactor code when told to. I can talk to it using language of the domain I just made up and I can talk to it using language that developers use when they talk about the code. It can solve bugs that would take me half an hour to figure out at least. It went from a poor junior to clever junior and half a senior in one year. It's not great at one shot architecting yet, but when told how to architect a thing it can move stuff around to conform with the desired architecture. It creates and modifies SVG's, understands colors and directions. It does all of that blind, without running the code. All it has is compiler (language server really) and linter. You mention it's an autocomplete. I barely used auto-complete at all. I just told it what to do. I touched the code a little bit only because I wanted to, never because I had to. When I wanted a change to the code, 95% of the time I just told it what change I needed and it made it. It basically replaces about 1/3 of software team already enabling (and pushing me towards) a role of (tech lead/architect/qa/product owner) and this fraction is rising. It's not flawless. Sometimes it makes things that don't work on the first try. But you don't always need to revert. It's capable of fixing what it made when told about how the desired behavior differs. And all that without running the program. I anticipate that further development is going to be letting it run and observe what it built so it's gonna move towards QA role. And the other thing might be slowly learning to differentiate between well architected code and badly architected code but that's probably harder as it requires careful preparation (and creation) of training data. While moving towards QA just requires figuring out how to let it run and inspect stuff. Maybe use the debugger.
  - bamboozled 1y ago
    Are you familiar with the garnet hype cycle? https://www.gartner.com/en/research/methodologies/gartner-hype-cycle https://www.gartner.com/en/research/methodologies/gartner-hy... I'm doing everything you're doing and probably more with Claude code (I 'let it run'). I'd say I too have been through a few of the different phases of the gartner hype cycle now. As I said it's a good tool, it's far from being an "auto-pilot".
    - scotty79 1y ago
      I don't care about the hype. If you put me in a cave for last 5 years and today gave me Windsurf and told me nothing beyond that a machine does that, I wouldn't believe you. All of it was literally science fiction 5 years ago. This literally replaces a person that I would have to hire to pursue my ideas.
      - bamboozled 1y ago
        Ok I guess that's where we differ, I don't really find it to be science fiction, the more I use it the system, the more I understand that magic of it, what it's going to do, and I've learned that to get good results, you absolutely have to drive it like almost every other tool. I guess if anyone went into a cave and came out for 5 years things would be like science fiction. FPV drones on the battlefield would be another example... As I said earlier, not downplaying it, really just warning you that if you lean to heavy on it without guiding it quite precisely, you will get burned, properly. All good, we have different opinions.
        scotty79 1y ago
        FPV drones on the battlefield were absolutely obvious. The only strange thing about it was, and still is, that powerful armies of the biggest players were "nah, I'll pass" on this obvious application. Wedding photographer figured out that you can hit people with drones. Farmers figured out they can carry and release significant payloads. Yet for some reason, generals seen them as toys and preferred to stick with their bombs and artillery shells costing order of magnitude more. We needed a war and soldiers on the contact line to basically start doing this obvious thing out of necessity and demanding supply from commanders. Computers being able to keep the conversation going about complex topic like development of a computer program were 5 years ago farther away than fusion energy. Perpetually 30 to 100 years away. Sure, for the best results you need to guide it because without it you'll get some program but not a good program. If you don't help it to control the complexity of the code with sensible architecture it will paint itself into a corner. But a junior developer would also do that if you just piled up requirements and never steered his output. I agree with you that it's a tool. But more like a combine harvester, that you still need to drive, for now, than a hammer. Already more like a CNC machine than a power drill. And still getting better. Everyone can have their own opinion on this. But I think next 5 years will change those opinions. For me it is at least as magical as cell phones. When I was a kid the peek of wireless personal communication was walkie-talkie with a very limited range. The idea that you could cram enough electronics, RF stuff and a battery in a small enough package and route the calls globally with vast networks of two way communication stations seemed absolutely ludicrous. I haven't even seen SMD element before I had my first cellphone and I picked apart a lot of electronics. Then it happened. It was the kind of technology that got me completely surprised, like LLMs. I remember not being that shocked because I was young and was expecting new things. I was reading a lot of SF and there instant wireless communication was everywhere. But after 4 decades of basically no progress in the machines that do anything that resembles thinking this innovation caught me off guard completely. I would be less surprised if we already had fusion reactors providing most of our electricity. On a more practical note, do you switch a lot the models that you use? I can't really tell them apart because I haven't worked with them all that much but I heard good things about GPT5 and when I used it it was very good (but others, like Claude or swt-1 were great too).
laserlight 1y ago
> it's hard to argue that there is no productivity boost I encourage you to keep up with the literature [0]. [0] https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/ https://metr.org/blog/2025-07-10-early-2025-ai-experienced-o...

jama211 1y ago

Personally I now get my work done in half the time and spend the rest of my newfound time with my family. It’s gonna be hard to convince me that isn’t a fantastic productivity increase.

eru 1y ago
That's interesting! Historically (at least for the last century or so) we mostly stopped lowering working hours and instead focused on increased output. Eg people still work eight hours a day, but get paid vastly more in real terms than in the 1950s. Instead of working only one hour a day (or fewer years or whatever) and taking a 1950s compensation package. Of course, this is merely a statistical observation. There are plenty of people who eg decide to retire early on a modest nest egg.
- actionfromafar 1y ago
  Funny how a 1950s compensation package could pay for a house and family though.
  - eru 1y ago
    1950s houses are no longer legal in most places. Both because in the US new construction is basically outlawed. But also because 1950s style houses and goods in general are so shoddy, you can no longer legally produce or buy them. In many cases, no one would even take them off your for free, even if they were still legal.) On a global scale, the times between perhaps 1940s to 1970s had some of the harshest inequality ever. It's been since about the 1980s that inequality has gone down markedly. Remember how India and China used to be on the verge of famine (or outright in famine). Nowadays obesity is the bigger problem.
- jama211 1y ago
  Remote work has enabled hours to be lowered again for many people, it’s just no one wants to say the quiet part out loud about that.
  - 1y ago

OtherShrezzing 1y ago

> There's zero evidence that LLMs have produced a measurable productivity boom I’m not so sure on this one. AlphaFold2, while not quite an LLM, was based on Transformer architecture - and its implementation wasn’t a million miles away from a language model - and it massively improved the rate of protein structure prediction. I think in general you’re correct, that we’re in a bubble, but I think it’s too extreme to say the technology is valueless.

JohnKemeny 1y ago
From what I've read, AlphaFold2 is a minor improvement over other solvers, and only in some cases. It makes many more mistakes and misleading results than classic solvers. But don't take my word for it, it's way outside of my area of research.

eru 1y ago

> And let's be real: we're in a bubble. There's zero evidence that LLMs have produced a measurable productivity boom. [...] You are right that a general productivity boom is hard to detect so far. But we already have certain industries with more than measurable productivity impact. A very salient example: first level call centre agents. The kind of unfortunate souls who are paid to man a phone and not allowed to deviate from their script, nor authorised to actually solve your problem. The companies renting out these services make heavy use of current AI and that has measurably affected hiring, staffing, prices etc. Even fairly basic AI (like what we have today) is good for this use case. Does it suck: sure. Does it suck more than the status quo with warm bodies without power: not really. And it's cheaper.

tim333 1y ago

"AI" is a marketing term... It's also been a coherent academic discipline since at least Turing's 1950 paper introducing the Turing test and with a history going back well beyond that.