51–60 of 87 posts
"Fourth, we will not find ourselves working less. Rather, there will be a bimodal outcome: those unable to find a place in the new AI-powered industry will fall out of it entirely, while those that remain are worked harder and harder as they drive automated development systems. Five PRs a week? Hah! Try fifty. Or five hundred." My experience with AI/LLMs summed up. The baseline expectation got higher. I didn't get an…
Earlier quoted context omitted.
To what end? Can't you just quit and get the same result?
Quitting is an individual action with potential adverse consequences (misery). The political question is why despite productivity in economic terms (which i know is flawed) growing many times over, do we still have to work as much, get paid so little, and have so many unemployed people looking for a job? Looks like without a parasitic capitalist class, we could share resources and work and have people live better liv…
"Sharing resources" doesn't work. We share way too much just to keep pensioners alive and can't have kids because of that. Confiscating all the billionaire's wealth wouldn't make a dent and would destroy much more.
Not sorry for the rant
Earlier quoted context omitted.
No, the more appropriate analogy would be that LLMs (reasoning agents would be better term, as they are no longer just an LLMs) are new beings rather than new tools. A new tool can be ultimately understood, and has a well-defined behaviour you can rely on. Then you can indeed become an expert in using a new tool. But AI agent is more a new being (or person). It's not possible to understand them, as they are not meant…
Most of the interaction surface with a coding agent is the harness anyway, so I'd definitely classify them as tools. Moreover, I don't like the anthropomorphism of calling them beings —and even if I were to follow that narrative, you can absolutely become better at interacting with people by practicing it. Regardless, it sure seems like developers experienced with using coding agents achieve better results than those…
I am not anthroporphizing them. I am just saying people are a better analogy.
> seems like developers experienced with using coding agents achieve better results than those without
That doesn't mean there is a skill involved. The same goes with people - you might just be naturally charismatic and get stuff done better with people. Doesn't mean it's a learnable skill that applies to everyone and every use case.
"Fourth, we will not find ourselves working less. Rather, there will be a bimodal outcome: those unable to find a place in the new AI-powered industry will fall out of it entirely, while those that remain are worked harder and harder as they drive automated development systems. Five PRs a week? Hah! Try fifty. Or five hundred." My experience with AI/LLMs summed up. The baseline expectation got higher. I didn't get an…
As it was the case with all productivity catalysts in the past. The Industrial Revolution promised machines would work and we’d get to have leisure time. Instead, it brought 16 hour factory shifts. This will be no different
Great piece, well written and succinctly sums up my thoughts. The bit I still don’t understand is how we all put up with the hallucinations. I was questioning Gemini last night about whether it could analyse a Fourtet song and give me a break down of the structure from beginning to end. “Sure!” it said with the endless enthusiasm you get from Gen tools, and then proceeded to spit out an absolute sack of fabricated sh…
> We’re so so far away from tools here that are anywhere near being trustworthy and accurate. And yet we (including myself) are chunking out code after code. It’s so bizarre. I think one more thing this whole LLM charade in the last few years has revealed is that no-one really cares. As long as it "looks" like it works, turns out, its all fine.
Great piece, well written and succinctly sums up my thoughts. The bit I still don’t understand is how we all put up with the hallucinations. I was questioning Gemini last night about whether it could analyse a Fourtet song and give me a break down of the structure from beginning to end. “Sure!” it said with the endless enthusiasm you get from Gen tools, and then proceeded to spit out an absolute sack of fabricated sh…
Doesn’t the article make the argument that since you can write tests this is not as much of a problem for code gen ? Its arguable whether it is a foolproof solution (I don’t think so) but it definitely makes it look like you can build a harness around the stochastic machine that will validate the correctness of the generated randomness. Monkeys and typewriters when you can quickly validate whether it’s Shakespeare or…
Great piece, well written and succinctly sums up my thoughts. The bit I still don’t understand is how we all put up with the hallucinations. I was questioning Gemini last night about whether it could analyse a Fourtet song and give me a break down of the structure from beginning to end. “Sure!” it said with the endless enthusiasm you get from Gen tools, and then proceeded to spit out an absolute sack of fabricated sh…
> We’re so so far away from tools here that are anywhere near being trustworthy and accurate. And yet we (including myself) are chunking out code after code. It’s so bizarre. I think one more thing this whole LLM charade in the last few years has revealed is that no-one really cares. As long as it "looks" like it works, turns out, its all fine.
Bic pens. Disposable razors. Whipped cream in a can.
Add LLM code to the list.
Earlier quoted context omitted.
As it was the case with all productivity catalysts in the past. The Industrial Revolution promised machines would work and we’d get to have leisure time. Instead, it brought 16 hour factory shifts. This will be no different
> Instead, it brought 16 hour factory shifts. To paraphrase Marx, "And so, the equivalent amount of labour required to produce the goods needed to sustain a worker for a day is, say, 4 hours. But that doesn't mean the selfsame worker can work for no longer that 4 hours! He can be forced to work six, or eight, or twelve hours a day, and whatever additional goods he produces — that's the surplus product, which in this…
Marx is great at building a narrative that generates resentment if you buy his frame. But you don't need to buy his frame, and if you don't, you suffer a lot less resentment. It's no way to live.
Earlier quoted context omitted.
As it was the case with all productivity catalysts in the past. The Industrial Revolution promised machines would work and we’d get to have leisure time. Instead, it brought 16 hour factory shifts. This will be no different
The industrial revolution hasn’t led to people working less, but it has led to more output: created more luxuries, made previous-luxuries available to lower classes, and reduced extreme poverty.
Likewise, AI will not lead to a future where the machines are working for us while we can enjoy our free time; it’s just increasing the required output from AI-augmented workers. That will probably increase economic output, but like before not leisure time.
Earlier quoted context omitted.
> We’re so so far away from tools here that are anywhere near being trustworthy and accurate. And yet we (including myself) are chunking out code after code. It’s so bizarre. I think one more thing this whole LLM charade in the last few years has revealed is that no-one really cares. As long as it "looks" like it works, turns out, its all fine.
Convenience trumps quality. Bic pens. Disposable razors. Whipped cream in a can. Add LLM code to the list.
But I am not using more software. I mean their source codes might have gotten larger, but the count of tools/services I use is basically the same.
So this feels more like giving up nice handcrafted fountain pens for bic pens. But I am still using a couple pens overall. So no added convenience, just shittier quality.