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

I don't care how well your "AI" works

fokus.cool

Post reply on HN
Print view View original

661–670 of 808 posts

Re: I don't care how well your "AI" works

#661

Post by rozap »

yawnxyz wrote: post #567
rozap wrote: post #500

Earlier quoted context omitted.

Now tell the AI to distill a bunch of user goals into a living system which has to evolve over time, integrate with other systems, etc etc. And deliver and support that system I use Claude code every day and it is a slam dunk for situations like the one above, fiddly UIs and the like. Seriously , some of the best money I spend. But it is not good at more abstract stuff. Still a massive time saver for me and does effe…

right, but this is akin to arguing why the table saw also does not do x/y/z — I don't know why we only complain about AI and how it does NOT do everything well yet . Maybe it's expectations set by all the AI companies, idk, but this kind of mentality seems very particular to AI products and nothing else.

I'm not complaining about it, I said in my post that it's a huge time saver. It's here to stay, and that's pretty clear to see. It has mostly automated away the need for junior engineers, which just 5 years ago would have been a very unexpected outcome, but it's kind of the reality now.

All that being said:

There's a segment of the software eng population that has their heads in the sand about it and the argument basically boils down to "AI bad". Those people are in trouble because they are also the people who insist on a whole committee meeting and trail of design documents to change the color of a button on a website that sells shoes. Most of their actual hard skills are pretty easy to outsource to an AI.

There's also a techbro segment of the population, who are selling snake oil about AGI being imminent, so fire your whole team and hire me in order to outsource your entire product to an army of AI agents. Their thoughts basically boil down to "I'm a grifter, and I smell money". Nevermind the fact that the outcome of such a program would be a smoldering tire fire, they'll be onto the next grift by then.

As with literally everything, there are loud, crazy people on either side and the truth is in the middle somewhere.

Re: I don't care how well your "AI" works

#662

Post by syllogism »

pirates wrote: post #598
evnp wrote: post #435

Earlier quoted context omitted.

I'm interested in this. Code review, most egregiously where the "author" neglected to review the LLM output themselves, seems like a clear instance. What are some other examples? Something that should go in a "survival guide" for devs that still prefer to code themselves.

AI-generated docs, charts, READMEs, TOE diagrams. My company’s Confluence is flooded with half assed documentation from several different dev teams that either loosely matches or doesn’t match at all the behavior or configuration of their apps. For example they ask to have networking configs put into place and point us at these docs that are not accurate and then they expect that we’ll troubleshoot and figure out wha…

If I were a CTO or VP these days I think I'd push for a blanket ban on committing docs/readmes/diagrams etc along with the initial work. Teams can push stuff to a `slop/` folder but don't call it docs.

If you push all that stuff at the same time, it's really easy to get away with this soft lie, "job done". They can claim they thought it was okay and it was just an honest mistake there were problems. They can lie about how much work they really did.

READMEs or diagrams that are plans for the functionality are fine. Docs that describe finished functionality are fine. Slop that dresses up unfinished work as finished work just fucks everything up, and the incentives are misaligned so everyone's doing this.

Re: I don't care how well your "AI" works

#663

Post by stocksinsmocks »

lxgr wrote: post #193

> I find it particularly disillusioning to realize how deep the LLM brainworm is able to eat itself even into progressive hacker circles. Anything worth reading beyond this transparent and hopefully unsuccessful appeal to tribalism? Hackers have always tried out new technologies to see how they work – or break – so why would LLMs be any different? > the devaluation of our craft, in a way and rate we never anticipated…

There seems to be a surprising theme that programmers wages are some kind of sunk cost and that their work was free of defects before.

I can get bad code written for the cost of electricity now

Re: I don't care how well your "AI" works

#664

Post by touisteur »

yawnxyz wrote: post #567
rozap wrote: post #500

Earlier quoted context omitted.

Now tell the AI to distill a bunch of user goals into a living system which has to evolve over time, integrate with other systems, etc etc. And deliver and support that system I use Claude code every day and it is a slam dunk for situations like the one above, fiddly UIs and the like. Seriously , some of the best money I spend. But it is not good at more abstract stuff. Still a massive time saver for me and does effe…

right, but this is akin to arguing why the table saw also does not do x/y/z — I don't know why we only complain about AI and how it does NOT do everything well yet . Maybe it's expectations set by all the AI companies, idk, but this kind of mentality seems very particular to AI products and nothing else.

I'm OK pondering the right use for the tool for as long as it'll take for the dust to settle. And I'm OK too trying some of it myself. What I resent is the pervasive request/pressure to use it everywhere right now, or 'be left out'.

My biggest gripe with the hype, as there's so much talk of craftmanship here, is: most programmers I've met hate doing code reviews and a good proportion prefer rewriting to reading and understanding other people's code. Now suddenly everyone is to be a prompter and astute reviewer of a flood of code they didn't write and now that you have the tool you should be faster faster faster or there's a problem with you.

Re: I don't care how well your "AI" works

#665

Post by potbelly83 »

raw_anon_1111 wrote: post #641
potbelly83 wrote: post #635

Earlier quoted context omitted.

I don't know, look at someone like https://news.ycombinator.com/user?id=dmbaggett he seems to be an entrepreneur who enjoys what he's doing.

Now compare that to these founders. https://docs.google.com/spreadsheets/d/1Uy2aWoeRZopMIaXXxY2E... How many of them do you think started their companies out of “passion”? Some of the ones I spotted checked had a couple of non technical founders looking for a “founding engineer” that they could underpay with the promise of “equity” that would probably be worthless.

I'm not disagreeing with the fact that there's a shit ton of founders out there looking for a quick pay day (I'd guess the majority fall into that category). Just pointing out there are exceptions, and the exceptions can be quite successful.

Re: I don't care how well your "AI" works

#666

Post by oytis »

easterncalculus wrote: post #244

> I find it particularly disillusioning to realize how deep the LLM brainworm is able to eat itself even into progressive hacker circles. That's the thing, hacker circles didn't always have this 'progressive' luddite mentality. This is the culture that replaced hacker culture. I don't like AI, generally. I am skeptical of corporate influence, I doubt AI 2027 and so-called 'AGI'. I'm certain we'll be "five years away"…

IDK, to me it looks that hacker culture has always been progressive, it's just definition of what is progressive has changed somewhat.

But hacker culture always sought to empower an individual (especially a smart, tech-savvy individual) against corporations, and rejection of gen AI seems reasonable in this light.

If hacker culture wasn't luddite, it's because of the widespread belief that the new digital technology does empower the individual. It's very hard to believe the same about LLMs, unless your salary depends on it

Re: I don't care how well your "AI" works

#667

Post by tsoukase »

In graphics there is the uncanny valley effect: when the object approaches reality the experience degrades. A similar mode holds for AI: the more the agent resembles human thinking, feeling and (in the future) touch, the more distress creates. Because it is not and probably never be real.

Re: I don't care how well your "AI" works

#668

Post by cpburns2009 »

embedding-shape wrote: post #638
cpburns2009 wrote: post #157

Earlier quoted context omitted.

I'm a native English speaker. Your understanding and usage of the word "reliability" is correct, and that's the exact word I'd use in this conversation. The GP is playing a pointless semantics game.

It's not semantics, if the definition is "it does what it’s supposed to do" then probably all of the currently deployed LLMs are reliable according to that definition.

> "it does what it’s supposed to do"

That's the crux of the problem. Many proponents of LLMs over promise the capabilities, and then deny the underperformance through semantics. LLMs are "reliable" only if you're talking about the algorithms behind the scene and you ignore the marketing. Going off the marketing they are unreliable, incorrect, and do not do what they're "supposed to do".

Re: I don't care how well your "AI" works

#669

Post by fragmede »

carlosjobim wrote: post #475
fileeditview wrote: post #325

Earlier quoted context omitted.

I don't think programmers are the issue here. What you describe sounds to me more like the typical product management in a company. Stuff features into the thing until it bursts of bugs and is barely maintainable. I would love to do something like what you describe. Build a simple but solid and very specialized solution. However I am not sure there is demand or if I have the right ideas for what to do. You mention in…

There's a whole bunch of apps for invoicing, but if you try them, you'll see that they are excessively complicated. Probably because they want to cover all bases of all use cases. Meaning they aren't great for any use case. Like you say. The invoicing app in particular I was referring to is Cakedesk. Made by a solo developer who sells it for a fair price. Easy to use and has all the necessary functions. Probably the…

One person's "excessively complicated" is another person's "lackluster and useless" because it doesn't have enough features.

Re: I don't care how well your "AI" works

#670

Post by jdiff »

ihaveajob wrote: post #571
LPisGood wrote: post #518

Earlier quoted context omitted.

> hacker circles didn't always have this 'progressive' luddite mentality Richard Stallman has his email printed out on paper for him to read, and he only connects to the internet by using wget to fetch web pages and then has them printed off.

You're probably thinking about Donald Knuth, not Stallman. https://www-cs-faculty.stanford.edu/~knuth/email.html

Stallman does similar, just without the printing step, and he checks his email entirely within emacs.

https://stallman.org/stallman-computing.html

Post reply on HN