Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Firing programmers for AI is a mistake

defragzone.substack.com

Post reply on HN
Print view View original

91–100 of 886 posts

Re: Firing programmers for AI is a mistake

#91

Post by bloomingkales »

adamors wrote: post #21
ponector wrote: post #10

>>companies aren’t investing in junior developers It was a case before AI as well. Overall it reads the same as "Twitter will be destroyed by mass layoffs". But it is still online

Twitter is essentially losing money even with a bare bones staff https://www.theverge.com/2025/1/24/24351317/elon-musk-x-twit... It's being kept online because it's a good propaganda tool, not due to how it performs on the free market.

It's arguably depreciating in value faster than a new car. One of Elmo's worst judgement calls (and that's saying a lot). Altman jabbed at Elmo and offered 9billion for X, 1/4th the price Elmo paid.

It's kind of hilarious watching the piranhas go at each other:

https://finance.yahoo.com/news/elon-musk-reportedly-offers-9...

Re: Firing programmers for AI is a mistake

#92

Post by netcan »

So... to discuss this for real we should first admit how things look below the super-premium level.

Software engineering at Stripe, R&D at Meta and such... these are one end of a spectrum.

At the middle of the spectrum is a team spending 6 years on a bullshit "cloud platform strategy" for a 90s UI that monitors equipment at a factory and produces reports required for compliance.

A lot of these struggle to get anything at all done.

Re: Firing programmers for AI is a mistake

#93

Post by lnrd »

netcan wrote: post #67

Our ability to predict technological "replacement" is pretty shoddy. Take banking for example. ATMs are literally called "teller machines." Internet banking is a way of "automating banking." Besides those, every administrative aspect of banking went from paper to computer. Do banks employ fewer people? Is it a smaller industry? No. Banks grew steadily over these decades. It's actually shocking how little network enab…

Is there data about this or is just your perception? Because my perception would be different, for example in my country countless bank branches closed and a lot of banking jobs do not exist anymore thanks to widespread home-banking usage (which I also know differs from country to country). This is also from the tales of people that had careers in banking and now tell how less banking jobs there are compared to when they joined in the 80s.

I wouldn't be sure that growth as an industry/business is correlated to a growth in jobs too.

Maybe I'm wrong, I would love to see some data about it.

Re: Firing programmers for AI is a mistake

#94

Post by baq »

aleph_minus_one wrote: post #66
baq wrote: post #39

Earlier quoted context omitted.

Once AI achieves runaway self improvement predicting the future is even more pointless than it is today. You’re looking at an economy in which the best human is worse at any and all jobs than the worst robot. There are no past examples to extrapolate from.

> Once AI achieves runaway self improvement predicting the future is even more pointless than it is today. You’re looking at an economy in which the best human is worse at any and all jobs than the worst robot. There are no past examples to extrapolate from. You take these strange dystopian science-fiction stories that AI bros invent to scam investors for their money far too seriously.

Humans are notoriously bad at extrapolating exponentials.

Re: Firing programmers for AI is a mistake

#95

Post by soco »

bigfatkitten wrote: post #19

For each programmer who actually spends their time on complex design work or fixing difficult bugs, there are many more doing what amounts to clerical work. Adding a new form here, fiddling with a layout there. It is the latter class who are in real danger.

But the question is, are we there yet? I have yet to hear of an AI bot who can eat up a requirement and add said new form in the right place, or fiddle with a layout. Do you know any? So all those big promises you read right now are outright lies. When will we reach that point? I don't do gambling. But we are not there, regardless what the salespeople or fancy journalists might be claiming all day long.

Re: Firing programmers for AI is a mistake

#96

Post by aleph_minus_one »

baq wrote: post #94
aleph_minus_one wrote: post #66

Earlier quoted context omitted.

> Once AI achieves runaway self improvement predicting the future is even more pointless than it is today. You’re looking at an economy in which the best human is worse at any and all jobs than the worst robot. There are no past examples to extrapolate from. You take these strange dystopian science-fiction stories that AI bros invent to scam investors for their money far too seriously.

Humans are notoriously bad at extrapolating exponentials.

... and many people who make this claim are notoriously prone to extrapolating exponential trends into a far longer future than the exponential trend model is suitable for.

Addendum: Extrapolating exponentials is actually very easy for humans: just plot the y axis on a logarithmic scale and draw a "plausible looking line" in the diagram. :-)

Re: Firing programmers for AI is a mistake

#97

Post by HqatsR »

pyrale wrote: post #11

We have fired all our programmers. However, the AI is hard to work with, it expects specific wording in order to program our code as expected. We have hired people with expertise in the specific language needed to transmit our specifications to the AI with more precision.

Yes, the best way is to type the real program completely into the AI, so that ClosedAI gets new material to train on, the AI can make some dumb comments but the code works.

And the manager is happy that filthy programmers are "using" AI.

Re: Firing programmers for AI is a mistake

#98

Post by yodsanklai »

Draiken wrote: post #71
yodsanklai wrote: post #42

Earlier quoted context omitted.

Why half-assed developer? It could be high-level developer that take advantage of AI to be more productive. This will reduce team sizes.

Because a high level developer will still have to fix all the shit the AI gets wrong, and therefore won't be "2x more productive" like I read in many places. If they're that much better with AI, they were likely coding greenfield CRUD boilerplate that nobody uses anyways. When the AI generated crap is actually used, it becomes evident how bad it is. But yes, this will reduce team sizes regardless of it being good or…

I'd say this is a lot of wishful thinking. Personally, I know that I'm more productive with AI. In my personal projects, I can tackle bigger projects than what would have been possible otherwise.

Will that reduce the demand for programmers? I hope not, but it's plausible at least.

Re: Firing programmers for AI is a mistake

#99

Post by fakedang »

soco wrote: post #95
bigfatkitten wrote: post #19

For each programmer who actually spends their time on complex design work or fixing difficult bugs, there are many more doing what amounts to clerical work. Adding a new form here, fiddling with a layout there. It is the latter class who are in real danger.

But the question is, are we there yet? I have yet to hear of an AI bot who can eat up a requirement and add said new form in the right place, or fiddle with a layout. Do you know any? So all those big promises you read right now are outright lies. When will we reach that point? I don't do gambling. But we are not there, regardless what the salespeople or fancy journalists might be claiming all day long.

Cursor already does the latter task pretty well, as I'm sure other AI agents already do. AI struggles only when it's something complex, like dealing with a geometric object, or plugging together infra, or programme logic.

Last year, I built a reasonably complicated e-commerce project wholly with AI, using the zod library and some pretty convoluted e-commerce logic. While it was a struggle, I was able to build it out in a couple of weeks. And I had zero prior experience even building forms in react, forget using zod.

Now shipping it to production? That's something AI will struggle at, but humans also struggle at that :(

Re: Firing programmers for AI is a mistake

#100

Post by baq »

aleph_minus_one wrote: post #96
baq wrote: post #94

Earlier quoted context omitted.

Humans are notoriously bad at extrapolating exponentials.

... and many people who make this claim are notoriously prone to extrapolating exponential trends into a far longer future than the exponential trend model is suitable for. Addendum: Extrapolating exponentials is actually very easy for humans: just plot the y axis on a logarithmic scale and draw a "plausible looking line" in the diagram. :-)

ah the 'everything is linear on a log-log plot when drawn with a fat marker' argument :)
Post reply on HN