Comment by rickydroll
3 hours ago
I agree with you that it probably won't replace jobs, but it will let you get more done in less time. I suspect, based on some of the stuff I've read recently, that two days with AI assistance are as exhausting/wearing as five days without, because you're trying to keep up with AI speed (another argument for slower LLMs on a local machine).
Unfortunately, I think we will also see corporate expectations that if you can get your previously normal amount of work done in two days, commonly you should be given more work. They don't need to hire that person or a replacement person for the disloyal/weak person who bailed, claiming they were overloaded.
It actually leads to more done in more time at the moment. For things that carry actual value at least.
For example your AI generated code feels very productive and "more done in less time", until you have to spend 3 times as much reviewing it, and supporting it long term compared to code written before AI.
People mistake instant gratification for long term productivity gains.
Well, if you throw a long-winded prompt over a fence and tell the agents to work it out on their own, yeah, you will spend a lot of time reviewing it. But if you use it as a form of pair programming and keep your scope of focus constrained, review time is built into your effort.
Also, keep in mind that my perspective is colored by the fact that I can't type much, and coding by keyboard is pretty much off the table for me.
And last, as I pointed out elsewhere, LLMs are useful for more things than writing code, and administrative work can be handled easily with the current generation of LLMs.
This is still a nonsense argument.
Do more of.. what? Much of what happens in software is not akin to a factory producing goods.
You need projects - projects don’t fall out the sky and value creating ones are hard to come by!
The funny thing about that is, I would argue, the opposite. Much of what happens in software is like a factory producing goods. I've worked on many projects at many companies; 80% of the code was written some variation of what I'd written elsewhere in different languages. We had to slog through the boring, "Been there, done that" code to get to the 20% that was interesting.
Also consider that there's a lot going on in the world that doesn't involve software creation. My partner works at a well-known law school, and a small but significant part of her work could be handled by AI. I'm looking into taking the department-specific info and RAGging it into an LLM so that students and instructors can find what they need in a chat interface, rather than using her as an LLM.