Live data from Hacker News

AI coding

geohot.github.io

241–250 of 299 posts

Re: AI coding

#241
I really hate the comparison of AI to a compiler. It's not like I'm writing a codebase in English. I'm still writing codebases in programming languages. AI just helps manage all the text.

Re: AI coding

#242

I'm 72, a dev for 40 years. I've lost a step or two. It's harder to buckle down and focus, but using AI tools have enabled me to keep building stuff. I can spec a project, have an agent build it then make sure it works. I just code for fun anyway.

I love using AI to prototype that something is possible then go and build it myself while borrowing bits from that initial MVP. The other night I wanted to build a browser extension that could intercept requests from a tab, claude got me something working in about 10 minutes with a couple prompts and a local storage session and then I toyed with the UI a bit to see what was possible.

Now after a weekend morning I have something much slimmer, predictable and sophisticated running... my extension shows a list of repeated responses and I can toggle which one to send to a localhost api that has a simple job queue to update a sqlite db with each new entry, extract the important parts and send it to my lm studio gpt oss 20b endpoint for some analysis and finally and send me a summary on telegram.

I know what I want in my head but cutting down the experimenting or PoC step down to minutes vs hours is pretty useful and as a competent enough dev it's elevated what I can get done now so I can take on more work than I would have by myself previously.

Re: AI coding

#243

> AI makes you feel 20% more productive but in reality makes you 19% slower. How many more billions are we going to waste on this? Adderall is similar. It makes people feel a lot more productive, but research on its effectiveness[0] seems to show that, at best, we get only a mild improvement in productivity, and marked deterioration of cognitive abilities. [0] https://pmc.ncbi.nlm.nih.gov/articles/PMC6165228/

I suppose you've never used Adderall during coding sessions before (and assuming you don't have a prescription for ADHD).

I used it when it was called Ritalin. It interacts with alcohol, too. Interesting stuff (not exactly the same stuff, but damn similar).

A lot of it. Other stuff, too, that makes it look like a cup of weak tea, but it’s been 45 years.

I’m pretty familiar with the illusion of productivity.

Re: AI coding

#244
Anyone dismissing AI coding without having seriously tried it or looked at the outputs of people who have can't be taken seriously. Yes, it shouldn't worked in theory. But it's disturbingly good in practice, for someone who knows what they are doing.

Re: AI coding

#245

I'm almost 50, and have been writing code professionally since the late 90s. I can pretty much see projects in my head, and know exactly what to build. I also get paid pretty well for what I do. You'd think I'd be the prototype for anti-AI. I'm not. I can build anything, but often struggle with getting bogged down with all the basic work. I love AI for speed running through all the boring stuff and getting to the goo…

I'm just slightly younger than you, but have the exact same sentiment. Hell, even moreso maybe, because what I realized is that "writing code to implement interesting ideas" is not really what I enjoy - it's coming up with the interesting ideas and experimenting with them. I couldn't care less about writing the code - and I only did it because I had to...if I wanted to see my idea come to life.

AI has also been a really good brainstorming partner - especially if you prompt it to disable sycophancy. It will tell you straight up when you are over-engineering something.

It's also wonderful at debugging.

So I just talk to my computer, brainstorm architectures and approaches, create a spec, then let it implement it. If it was a bad idea, we iterate. The iteration loop is so fast that it doesn't matter.

Did you end up regretting a design choice, but normally you'd live with it because so much code would have to be changed? Not with Agentic coding tools - they are great at implementing changes throughout the entire codebase.

And its so easy to branch out to technologies you're not an expert in, and still be really effective as you gain that expertise.

I honestly couldn't be happier than I am right now. And the tools get better every week, sometimes a couple times a week.

Re: AI coding

#246

Earlier quoted context omitted.

No. I'm just saying: "yes, AI can code."

no one is disagreeing. you just havent discovered the true costs yet

what is the cost of enabling someone to create little pieces of software if he otherwise wouldn't? I'm no advocate of vibe coding serious commercial software, but to do little DIY stuff - absolutely.

Re: AI coding

#247

Anyone dismissing AI coding without having seriously tried it or looked at the outputs of people who have can't be taken seriously. Yes, it shouldn't worked in theory. But it's disturbingly good in practice, for someone who knows what they are doing.

So anyone that doesn't know what they're doing can dismiss it without seriously trying it? I mean, I agree, but I feel like you wouldn't with this implication of what you're saying...

Re: AI coding

#248

Earlier quoted context omitted.

no one is disagreeing. you just havent discovered the true costs yet

what is the cost of enabling someone to create little pieces of software if he otherwise wouldn't? I'm no advocate of vibe coding serious commercial software, but to do little DIY stuff - absolutely.

No one is forbidding people creating little pieces of software.

But creating little pieces of software was already available, for example I make most of my DIY software in a spreadsheet.

There are tons upon tons of low code possibilities or already existing software packages that need a bit of configuration that one can use and using AI or LLM's is not bringing anything that is revolutionizing access to computation or tailoring software. Just get your head around Excel or LibreOffice Calc and most of your DIY software needs are covered.

What LLM's and AI is bringing to the table is illusion of being able to create software exactly like people who have all the theoretical background or experience in building it. (which seems the person I originally replied to be one of those)

So as my original comment the problem is: that people building a forest hut are convinced that only if they just put more sticks on top, they will manage to build a skyscraper.

Re: AI coding

#249
post #48

I'm almost 50, and have been writing code professionally since the late 90s. I can pretty much see projects in my head, and know exactly what to build. I also get paid pretty well for what I do. You'd think I'd be the prototype for anti-AI. I'm not. I can build anything, but often struggle with getting bogged down with all the basic work. I love AI for speed running through all the boring stuff and getting to the goo…

I love AI for speed running through all the boring stuff and getting to the good parts. In some cases, especially with the more senior devs in my org, fear of the good parts is why they're against AI. Devs often want the inherent safety of the boring, easy stuff for a while. AI changes the job to be a constant struggle with hard problems. That isn't necessarily a good thing. If you're actually senior by virtue of tim…

This is an interesting take on it. I've been using Cursor at work and notice that days where I have it generate all the "easy parts", I end up feeling much more mentally exhausted than if I spend the day writing everything by hand.

Re: AI coding

#250

Earlier quoted context omitted.

The usefulness is in saving time boilerplating, plus figuring out tests I may not have thought of. But I do closely review the code! It turns the usual drudge of writing tests into more of a code review. Last time I did it it had some mistakes I needed to fix for sure.

There shouldn't be boilerplate in test code. It should be refactored into harnesses, utils, and fixtures instead.

While absolutist, I think this is a good goal! AI wont do that unprompted. It will get the job done TM. But not elegantly.

A lot of the tests have those things. Boilerplate becomes knowing which to use, as well as the it.each(...) etc. ceremonies

Post reply on HN