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

Andrej Karpathy: Software in the era of AI [video]

youtube.com

Post reply on HN
Print view View original

821–827 of 827 posts

Re: Andrej Karpathy: Software in the era of AI [video]

#821

Post by infecto »

bopbopbop7 wrote: post #817
infecto wrote: post #816

Earlier quoted context omitted.

It’s not obvious to me but perhaps you are approach it from a biased perspective. Sure if you left all higher order function to a LLM, thinking about homework and simply parsing it all through a chatbot, of course you are losing out. There is a lot of nuance to it and I am not sure if that very first initial study captures it. Everyone is different and YMMV but I suspect it will come down to how you use the tools not…

You asked for a study, you got a study. Yet you’re still playing mental gymnastics to somehow prove that not using your brain doesn’t impair thinking. You want an anecdote now? After dismissing a study? Wild. I doubt that will convince you if a study won’t, just grasping at straws. And no, it’s not nuanced at all. If you stop using your brain, you lose cognitive abilities. If you stop working out, you lose muscle. If…

You didn’t “call me out with evidence,” you linked a single early study that doesn’t prove the absolutist claim you’re making. You’re taking a broad, complex topic and flattening it into a gym analogy. That’s lazy thinking. It’s far too early to say either positions. I also never asked for a study, I asked for the parent to back up his claims that are far exceeding anything that study may show. He has nothing. While you might be unable to hold a constructive discussion, I will repeat mine.

I suspect there are shades of gray in how these tools are being used. From one extreme end of just using it to as a magic ball of decision making where no thinking goes into the process, to the other end where a constructive Q&A style discussion is being had. I would be surprised if having constructive discussion with a LLM to not be activating the brain, but like I said, I welcome more studies to dig more as what we have so far today is not what I would call conclusive or showing much.

Green accounts being green.

Re: Andrej Karpathy: Software in the era of AI [video]

#822

Post by bopbopbop7 »

infecto wrote: post #821
bopbopbop7 wrote: post #817

Earlier quoted context omitted.

You asked for a study, you got a study. Yet you’re still playing mental gymnastics to somehow prove that not using your brain doesn’t impair thinking. You want an anecdote now? After dismissing a study? Wild. I doubt that will convince you if a study won’t, just grasping at straws. And no, it’s not nuanced at all. If you stop using your brain, you lose cognitive abilities. If you stop working out, you lose muscle. If…

You didn’t “call me out with evidence,” you linked a single early study that doesn’t prove the absolutist claim you’re making. You’re taking a broad, complex topic and flattening it into a gym analogy. That’s lazy thinking. It’s far too early to say either positions. I also never asked for a study, I asked for the parent to back up his claims that are far exceeding anything that study may show. He has nothing. While…

[dead]

Re: Andrej Karpathy: Software in the era of AI [video]

#825

Post by yencabulator »

dist-epoch wrote: post #258
khalic wrote: post #253

Earlier quoted context omitted.

Phi4-mini runs on a basic laptop CPU at 20T/s… how is that slow? Without optimization…

I was running Qwen3-32B locally even faster, 70T/s, still way too slow for me. I'm generating thousands of tokens of output per request (not coding), running locally I could get 6 mil tokens per day and pay electricity, or I can get more tokens per day from Google Gemini 2.5 Flash for free. Running models locally is a privilege for the rich and those with too much disposable time.

Try Qwen3-30B-A3B. It's MoE to an extent where its use of memory bandwidth looks more like a 3B model, and thus it typically goes faster.

Re: Andrej Karpathy: Software in the era of AI [video]

#826

Post by chamomeal »

consumer451 wrote: post #617
diggan wrote: post #378

Earlier quoted context omitted.

> Does anybody really want to be an assembly line QA reviewer for an automated code factory? Sounds like shit. On the other hand, does anyone really wanna be a code-monkey implementing CRUD applications over and over by following product specifications by "product managers" that barely seem to understand the product they're "managing"? See, we can make bad faith arguments both ways, but what's the point?

I hesitate to divide a group as diverse as software devs into two categories, but here I go: I have a feeling that devs who love LLM coding tools are more product-driven than those who hate them. Put another way, maybe devs with their own product ideas love LLM coding tools, whilr devs without them do not. I am genuinely not trying to throw shade here in any way. Does this rough division ring true to anyone else? Is…

No I think that’s accurate! But maybe instead of “devs who think about product stuff vs devs who don’t”, it depends on what hat you’re wearing.

When I’m working on something that I just want it to work, I love using LLMs. Shell functions for me to stuff into my config and use without ever understanding, UI for side projects that I don’t particularly care about, boilerplate nestjs config crap. Anything where all I care about is the result, not the process or the extensibility of the code: I love LLMs for that stuff.

When it’s something that I’m going to continue working on for a while, or the whole point is the extensibility/cleanliness of the program, I don’t like to use LLMs nearly as much.

I think it might be because most codebases are built with two purposes: 1) to be used as a product 2) to be extended and turned into something else

LLMs are super good at the first purpose, but not so good at the second.

I heard an interesting interview on the playdate dev podcast by the guy who made Obra Dinn. He said something along the lines of “making a game is awesome because the code can be horrible. All that matters is that the game works and is fun, and then you are done. It can just be finished, and then the code quality doesn’t matter anymore.”

So maybe LLMs are just really good for when you need something specific to work, and the internals don’t matter too much. Which are more the values of a product manager than a developer.

So it makes sense that when you are thinking more product-oriented, LLMs are more appealing!

Re: Andrej Karpathy: Software in the era of AI [video]

#827

Post by alightsoul »

alightsoul wrote: post #79

why does vibe coding still involve any code at all? why can't an AI directly control the registers of a computer processor and graphics card, controlling a computer directly? why can't it draw on the screen directly, connected directly to the rows and columns of an LCD screen? what if an AI agent was implemented in hardware, with a processor for AI, a normal computer processor for logic, and a processor that correlat…

This is basically what Lovable is. It's a multi billion dollar company today
Post reply on HN