5 ms·

I didn’t see this at first, and I was fairly shaken by the potential impact on the world if their progress didn’t stop. A couple generations showed meaningful i

I didn’t see this at first, and I was fairly shaken by the potential impact on the world if their progress didn’t stop. A couple generations showed meaningful improvements, but now it seems like you’re probably correct. I’ve used these for years quite intensively to aid my work and while it’s a useful rubber duck, it doesn’t seem to yield much more beyond that. I worry a lot less about my career now. It really is a tool that creates more work for me rather than less.

iman453 2y ago
Would this still hold true in your opinion if models like O3 become super cheap and bit better over time? I don't know much about the AI space, but as a vanilla backend dev also worry about the future :)
- root_axis 2y ago
  Let's see how O3 pans out in practice before we start setting it as the standard for the future.
  - varelse 2y ago
    Mamba-ish models are the breakthrough to cheap inference if they pan out. Calling a dead-end already is just silly.
- sydd 2y ago
  We know that OpenAI is verz good at least in one thing: generating hype. When Sora was announced everyone thought that this will be revolutionary. Look at how it looks like in production. Same when they started floating rumours that they have some AGI prototype in their labs. They are the Tesla of the IT world, overpromise and under deliver.
  - WhyOhWhyQ 2y ago
    It's a brilliant marketing model. Humans are inherently highly interested in anything which could be a threat to their well-being. Everything they put out is a tacit promise that the viewer will soon be economically valueless.
    - bilbyx 2y ago
      I hope people will come to the realisation that we have created a good plagiarizer at best. The "intelligence" originates from the human beings who created the training data for these LLMs. The hype will die when reality hits.
  - gom_jabbar 2y ago
    Hype is very interesting. The concept of Hyperstition describes fictions that make themselves real. In this sense, hype is an essential part of capitalism: "Capitalization is [...] indistinguishable from a commercialization of potentials, through which modern history is slanted (teleoplexically) in the direction of ever greater virtualization, operationalizing science fiction scenarios as integral components of production systems." [0] "Within capitalist futures markets, the non-actual has effective currency. It is not an "imaginary" but an integral part of the virtual body of capital, an operationalized realization of the future." [1] This corresponds to the idea that virtual is opposed to actual, not real. [0] https://retrochronic.com/#teleoplexy-12 https://retrochronic.com/#teleoplexy-12 [1] https://retrochronic.com/#on-accelerate-2b https://retrochronic.com/#on-accelerate-2b
    - seadan83 2y ago
      Religion too. Those that are told a prophecy is to come, have a lot of incentive to fulfill that prophecy. Human belief systems are strange and interesting because (IMO) of the entanglement of beliefs with identity
- hatefulmoron 2y ago
  I'm really curious about something, and would love for an OpenAI subscriber to weigh in here. What is the jump to O1 like, compared to GPT4/Claude 3.5? I distinctly remember the same (if not even greater) buzz around the announcement of O1, but I don't hear people singing its praises in practice these days.
  - te_chris 2y ago
    O1 is fine.
  - lom888 2y ago
    I don't know how to code in any meaningful way. I work at a company where the bureaucracy is so thick that it is easier to use a web scraper to port a client's website blog than to just move the files over. GPT 4 couldn't write me a working scraper to do what I needed. o1 did it with minimal prodding. It then suggested and wrote me a ffmpeg front-end to handle certain repetitive tasks with client videos, again, with no problem. Gpt4 would often miss the mark and then write bad code when presented with such challenges
    - 2muchcoffeeman 2y ago
      Try Claude. I get even better code results.
  - jazzyjackson 2y ago
    I gave up interest in GPT4/Claude3.5 about 6 months ago as not very helpful, producing plausible but wrong code. Have an o3-mini model available to me on the other hand I'm very impressed with its fast, succinct, correct answers while tooling around in zsh on my mac. what things are called, why they exist. why is macports installing db48 etc. It still fails to write simple bash one liners. (I wanted to pipe the output of ffmpeg to a column of --enabled-features and it just couldn't do it) It's a very helpful rubber duck but still not going to suffice as an agent, but I think its worth a subscription. I wanted to do everything local and self hosted and briefly owned a $3000 mac studio to run llama3.3-70B but it was only as good as GPT4 and too slow to be useful so returned it. In that context even $200/m is relatively cheap.
- foobarian 2y ago
  I was helping a relative still in college with a project, and I was struck by how lackadaisical they are about cut-and-pasting huge chunks of code from chatgpt into whatever module they are building without thinking about why, or what it does, or where it fits, as long as it works. It doesn't help that it's all relatively same-looking Javascript so frontend or backend is kinda mixed together. The troubleshooting help I provided was basically untangling the mess by going from first principles and figuring out what goes where. I can tell you I did not feel threatened by the AI there at all, if anything I felt bad for the juniors and feeling like this is what we old people are going to end up having to support very soon.
  - muixoozie 2y ago
    I had a similar experience recently. Not sure how accurate these numbers are but on https://openrouter.ai/ https://openrouter.ai/ highest used "apps" basically can auto-accept generated code and apply it to the project. I was recently looking at top performers on https://www.swebench.com/ https://www.swebench.com/ and noticed OpenHands basically does the same thing or similar. I think the trend is going to get much worse, and I don't think Moore's Law is going to save us from the resulting chaos.
  - docmars 2y ago
    AI slop-fixing consultants, at your service! There's hope for us veterans yet. ;)
- intended 2y ago
  It will be kinda like modern furniture, vs “Old Human written code.” So many people won’t be able to get paid for the quality of their work, but the people who buy the cheaper product will get what they pay for.
- tempodox 2y ago
  No degree of cheapness will be able to offset the “creates more work for me rather than less” part.
- steve_adams_86 2y ago
  Generally speaking, I think it would. I’m open to being wrong. I think there is a non-trivial amount of hype around O3, and while it would certainly be interesting if it was cheap, I don’t think it would address important issues that AI currently doesn’t seem to even begin to accommodate in its current capacity to recognize or utilize contexts. For example, I have little to no expectation that it will handle software architecture well. Especially refactoring legacy code, where two enormous contexts need to be held in mind at once.
trhway 2y ago
>I worry a lot less about my career now. It really is a tool that creates more work for me rather than less. when i was a team/project leader the largest part of my work was talking to the reports on what needs to be implemented and how they are going to implement it and the current progress of the implementation, how to interface the stuff, what are the issues and how to approach the troubleshooting, what are the next steps, etc. with occasional looking into/reviewing the code - it looks to me what working with coding LLM would soon be quite similar to that.
trod1234 2y ago
Many of the major harms of these things were neglected, and downplayed even to this day people don't recognize just how changed the world has become. The mere delusion that AI will replace work has been used to justify mass layoffs. The persistence of indistinct ghost jobs that are generated by computer for pennies to flood and bind with prospective job seekers (similar to RNA interference), has resulted in severe brain drain in many fields. Worse, the fact these people have often been forced into poverty as a result will have a lasting impact. You might have planned for up to a year out of work pre-AI and had the financial resources, but now how long does it take? Conversion ratios for the first step have changed by two magnitudes or order (from x100 to x10,000). What are the odds of these people finding a job given their finite time, and requirements that are un-automatable for submission (nil). The media keeps claiming that everything is getting better, the stats say so (while neglecting the fact that the stats are being manipulated to the point of uselessness/fabricated), but you have 1/3 of welfare payouts now going to these people in California (in the US), just for basic food. When you can't find work, you go where the work is abandoning the bad economic investment and choice you made regardless of how competent you were. It is a psychologically sticky decision. When there is no chance at finding work, you get desperate, and many desperate people turn to crime and unrest. This was foreseen by a number of very intelligent people many decades ago, and ignored following business as usual. The mere demonstration that we are unable to react in time is what gave engineers such great pause to write about these things, as far back as in the 70s. Hysteresis is a lagging time problem where you can't react fast enough to avert catastrophic failure given chaotic conditions, leaving survival up to chance. Its the worst type of engineering problem with real consequences. Given how western society is structured dependently on labor exchange, its a perfect weapon of chaos and debasement in the value of labor, that effectively destroys half of its underlying economic structure (factor markets). This forces sieving conditions of wealth that become spinodal, and eventually falter under their constraints and spiral into deflationary trends over time. Business wins so much that they lose everything. Its quite a disadvantaged environment and the general trend is that everyone is ignoring the pink elephant. Actions (and inaction) have consequences. When people don't listen and take appropriate action, consequences get dire, it hits the fan.
- cgio 2y ago
  I agree that we are often missing in our analyses the true materialised impact of expectations by focusing on the validity of said expectations instead. Organisations, even if not laying off, are pausing hiring plans with a conviction that AI will replace some of the workers. It then becomes a self fulfilling prophecy to some extent. It doesn’t matter if it can, what matters is if it will. And to assume that people won’t place a bet is futile, as everyone does, and even if it’s wrong the market will allocate the losses to the baseline.

iman453 2y ago

Would this still hold true in your opinion if models like O3 become super cheap and bit better over time? I don't know much about the AI space, but as a vanilla backend dev also worry about the future :)

root_axis 2y ago
Let's see how O3 pans out in practice before we start setting it as the standard for the future.
- varelse 2y ago
  Mamba-ish models are the breakthrough to cheap inference if they pan out. Calling a dead-end already is just silly.
sydd 2y ago
We know that OpenAI is verz good at least in one thing: generating hype. When Sora was announced everyone thought that this will be revolutionary. Look at how it looks like in production. Same when they started floating rumours that they have some AGI prototype in their labs. They are the Tesla of the IT world, overpromise and under deliver.
- WhyOhWhyQ 2y ago
  It's a brilliant marketing model. Humans are inherently highly interested in anything which could be a threat to their well-being. Everything they put out is a tacit promise that the viewer will soon be economically valueless.
  - bilbyx 2y ago
    I hope people will come to the realisation that we have created a good plagiarizer at best. The "intelligence" originates from the human beings who created the training data for these LLMs. The hype will die when reality hits.
- gom_jabbar 2y ago
  Hype is very interesting. The concept of Hyperstition describes fictions that make themselves real. In this sense, hype is an essential part of capitalism: "Capitalization is [...] indistinguishable from a commercialization of potentials, through which modern history is slanted (teleoplexically) in the direction of ever greater virtualization, operationalizing science fiction scenarios as integral components of production systems." [0] "Within capitalist futures markets, the non-actual has effective currency. It is not an "imaginary" but an integral part of the virtual body of capital, an operationalized realization of the future." [1] This corresponds to the idea that virtual is opposed to actual, not real. [0] https://retrochronic.com/#teleoplexy-12 https://retrochronic.com/#teleoplexy-12 [1] https://retrochronic.com/#on-accelerate-2b https://retrochronic.com/#on-accelerate-2b
  - seadan83 2y ago
    Religion too. Those that are told a prophecy is to come, have a lot of incentive to fulfill that prophecy. Human belief systems are strange and interesting because (IMO) of the entanglement of beliefs with identity
hatefulmoron 2y ago
I'm really curious about something, and would love for an OpenAI subscriber to weigh in here. What is the jump to O1 like, compared to GPT4/Claude 3.5? I distinctly remember the same (if not even greater) buzz around the announcement of O1, but I don't hear people singing its praises in practice these days.
- te_chris 2y ago
  O1 is fine.
- lom888 2y ago
  I don't know how to code in any meaningful way. I work at a company where the bureaucracy is so thick that it is easier to use a web scraper to port a client's website blog than to just move the files over. GPT 4 couldn't write me a working scraper to do what I needed. o1 did it with minimal prodding. It then suggested and wrote me a ffmpeg front-end to handle certain repetitive tasks with client videos, again, with no problem. Gpt4 would often miss the mark and then write bad code when presented with such challenges
  - 2muchcoffeeman 2y ago
    Try Claude. I get even better code results.
- jazzyjackson 2y ago
  I gave up interest in GPT4/Claude3.5 about 6 months ago as not very helpful, producing plausible but wrong code. Have an o3-mini model available to me on the other hand I'm very impressed with its fast, succinct, correct answers while tooling around in zsh on my mac. what things are called, why they exist. why is macports installing db48 etc. It still fails to write simple bash one liners. (I wanted to pipe the output of ffmpeg to a column of --enabled-features and it just couldn't do it) It's a very helpful rubber duck but still not going to suffice as an agent, but I think its worth a subscription. I wanted to do everything local and self hosted and briefly owned a $3000 mac studio to run llama3.3-70B but it was only as good as GPT4 and too slow to be useful so returned it. In that context even $200/m is relatively cheap.
foobarian 2y ago
I was helping a relative still in college with a project, and I was struck by how lackadaisical they are about cut-and-pasting huge chunks of code from chatgpt into whatever module they are building without thinking about why, or what it does, or where it fits, as long as it works. It doesn't help that it's all relatively same-looking Javascript so frontend or backend is kinda mixed together. The troubleshooting help I provided was basically untangling the mess by going from first principles and figuring out what goes where. I can tell you I did not feel threatened by the AI there at all, if anything I felt bad for the juniors and feeling like this is what we old people are going to end up having to support very soon.
- muixoozie 2y ago
  I had a similar experience recently. Not sure how accurate these numbers are but on https://openrouter.ai/ https://openrouter.ai/ highest used "apps" basically can auto-accept generated code and apply it to the project. I was recently looking at top performers on https://www.swebench.com/ https://www.swebench.com/ and noticed OpenHands basically does the same thing or similar. I think the trend is going to get much worse, and I don't think Moore's Law is going to save us from the resulting chaos.
- docmars 2y ago
  AI slop-fixing consultants, at your service! There's hope for us veterans yet. ;)
intended 2y ago
It will be kinda like modern furniture, vs “Old Human written code.” So many people won’t be able to get paid for the quality of their work, but the people who buy the cheaper product will get what they pay for.
tempodox 2y ago
No degree of cheapness will be able to offset the “creates more work for me rather than less” part.
steve_adams_86 2y ago
Generally speaking, I think it would. I’m open to being wrong. I think there is a non-trivial amount of hype around O3, and while it would certainly be interesting if it was cheap, I don’t think it would address important issues that AI currently doesn’t seem to even begin to accommodate in its current capacity to recognize or utilize contexts. For example, I have little to no expectation that it will handle software architecture well. Especially refactoring legacy code, where two enormous contexts need to be held in mind at once.

trhway 2y ago

>I worry a lot less about my career now. It really is a tool that creates more work for me rather than less. when i was a team/project leader the largest part of my work was talking to the reports on what needs to be implemented and how they are going to implement it and the current progress of the implementation, how to interface the stuff, what are the issues and how to approach the troubleshooting, what are the next steps, etc. with occasional looking into/reviewing the code - it looks to me what working with coding LLM would soon be quite similar to that.

trod1234 2y ago

Many of the major harms of these things were neglected, and downplayed even to this day people don't recognize just how changed the world has become. The mere delusion that AI will replace work has been used to justify mass layoffs. The persistence of indistinct ghost jobs that are generated by computer for pennies to flood and bind with prospective job seekers (similar to RNA interference), has resulted in severe brain drain in many fields. Worse, the fact these people have often been forced into poverty as a result will have a lasting impact. You might have planned for up to a year out of work pre-AI and had the financial resources, but now how long does it take? Conversion ratios for the first step have changed by two magnitudes or order (from x100 to x10,000). What are the odds of these people finding a job given their finite time, and requirements that are un-automatable for submission (nil). The media keeps claiming that everything is getting better, the stats say so (while neglecting the fact that the stats are being manipulated to the point of uselessness/fabricated), but you have 1/3 of welfare payouts now going to these people in California (in the US), just for basic food. When you can't find work, you go where the work is abandoning the bad economic investment and choice you made regardless of how competent you were. It is a psychologically sticky decision. When there is no chance at finding work, you get desperate, and many desperate people turn to crime and unrest. This was foreseen by a number of very intelligent people many decades ago, and ignored following business as usual. The mere demonstration that we are unable to react in time is what gave engineers such great pause to write about these things, as far back as in the 70s. Hysteresis is a lagging time problem where you can't react fast enough to avert catastrophic failure given chaotic conditions, leaving survival up to chance. Its the worst type of engineering problem with real consequences. Given how western society is structured dependently on labor exchange, its a perfect weapon of chaos and debasement in the value of labor, that effectively destroys half of its underlying economic structure (factor markets). This forces sieving conditions of wealth that become spinodal, and eventually falter under their constraints and spiral into deflationary trends over time. Business wins so much that they lose everything. Its quite a disadvantaged environment and the general trend is that everyone is ignoring the pink elephant. Actions (and inaction) have consequences. When people don't listen and take appropriate action, consequences get dire, it hits the fan.

cgio 2y ago
I agree that we are often missing in our analyses the true materialised impact of expectations by focusing on the validity of said expectations instead. Organisations, even if not laying off, are pausing hiring plans with a conviction that AI will replace some of the workers. It then becomes a self fulfilling prophecy to some extent. It doesn’t matter if it can, what matters is if it will. And to assume that people won’t place a bet is futile, as everyone does, and even if it’s wrong the market will allocate the losses to the baseline.