[ben_w]

Depends what the problem is.

Sometimes you can, sometimes you have to break the problem apart and get the LLM to do each bit separately, sometimes the LLM goes funny and you need to solve it yourself.

Customers don't want you wasting money doing by hand what can be automated, nor do they want you ripping them off by blindly handing over unchecked LLM output when it can't be automated.

[agumonkey]

there are other ways: being scammed by lazy devs using AI to produce what devs normally do and not saving any money for the customer. i mentioned it in another thread, i heard first hand people say "i will never report how much time savings i get from gemini, at best i'll say 1 day a month"

[Dylan16807]

If you produce the same product, then you get to ask for the same pay. That's not a scam.

If enough people can make the product faster, then competition will drive the price down. But the ability to charge less is not at all an obligation to charge less.

[agumonkey]

it's paradoxical, the llm is not helping consumers, it's not helping the experienced engineer, it's helping a new class of devs that just want the easy way out. and ultimately this wave will make the price go down to the point the skilled dev won't be able to sustain long term growth because learning more and more advanced will not be valued by the economy.. just a thought but i don't see a nice path ahead now

[Dylan16807]

Why is it not helping the experienced engineer? I don't fully understand your scenario.

If the experienced engineer is already faster than the LLM, their job is not at risk.

If the LLM is faster then the experienced engineer at making some kind of code product, then the experienced engineer can use it to save time. And in the short term they can spend even more time learning! Maybe it's a net negative because it helps the "new class of devs that just want the easy way out" more, but it's still helping the experienced engineer.

And if increased competition drops the price then the LLM's influence is helping customers.

[agumonkey]

you may have a point, i'm fuzzy in my perception right now but there are non linear factor imo, here's how i see things

- a market needs a certain kind of product (feature set, complexity, performance)

- good engineer could apply skills to deliver that

- lazy engineers couldn't, but with llm they can, it gives them solutions without understanding much, which is irrelevant for them, they want to ship

- i myself don't enjoy having code spilled out for me, and the time savings from llm won't bring much more joy (unlike the lazy engineer who is happy)

- a llm might help me do more advanced things but the market might not care for it. say the average user wants a dashboard with a bunch of data points and a few actions. the llm answer will match that perfectly. I could ask the llm to produce a more complex dashboard with more customization, more feature.. but the user will not want it because it's beyond its needs

so yeah it's a matter of ratio, it seems that lazy devs will get a 10x improvement while a skilled one will only get 1.5 and might be squeezed out of the market