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

401–410 of 827 posts

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

#401

Post by thierrydamiba »

blixt wrote: post #394

If we extrapolate these points about building tools for AI and letting the AI turn prompts into code I can’t help but reach the conclusion that future programming languages and their runtimes will be heavily influenced by the strengths and weaknesses of LLMs. What would the code of an application look like if it was optimized to be efficiently used by LLMs and not humans? * While LLMs do heavily tend towards expectin…

Is a world driven by the strengths and weaknesses of programming languages better than the one driven by the strengths and weaknesses of LLMs?

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

#402

Post by mythrwy »

blixt wrote: post #394

If we extrapolate these points about building tools for AI and letting the AI turn prompts into code I can’t help but reach the conclusion that future programming languages and their runtimes will be heavily influenced by the strengths and weaknesses of LLMs. What would the code of an application look like if it was optimized to be efficiently used by LLMs and not humans? * While LLMs do heavily tend towards expectin…

It does seem a bit silly long term to have something like Python which was developed as a human friendly language written by LLMs.

If AI is going to write all the code going forward, we can probably dispense with the user friendly part and just make everything efficient as possible for machines.

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

#403

Post by BobbyJo »

karpathy wrote: post #22
gchamonlive wrote: post #2

I think it's interesting to juxtapose traditional coding, neural network weights and prompts because in many areas -- like the example of the self driving module having code being replaced by neural networks tuned to the target dataset representing the domain -- this will be quite useful. However I think it's important to make it clear that given the hardware constraints of many environments the applicability of what…

I kind of say it in words (agreeing with you) but I agree the versioning is a bit confusing analogy because it usually additionally implies some kind of improvement. When I’m just trying to distinguish them as very different software categories.

The versioning makes sense to me. Software has a cycle where a new tool is created to solve a problem, and the problem winds up being meaty enough, and the tool effective enough, that the exploration of the problem space the tool unlocks is essentially a new category/skill/whatever.

computers -> assembly -> HLL -> web -> cloud -> AI

Nothing on that list has disappeared, but the work has changed enough to warrant a few major versions imo.

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

#404

Post by stuartmemo »

hgl wrote: post #51

It’s fascinating to think about what true GUI for LLM could be like. It immediately makes me think a LLM that can generate a customized GUI for the topic at hand where you can interact with in a non-linear way.

It's probably Jira. https://medium.com/question-park/all-aboard-the-ai-train-b03...

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

#405

Post by tudorizer »

95% terrible expression of the landscape, 5% neatly dumbed down analogies.

English is a terrible language for deterministic outcomes in complex/complicated systems. Vibe coders won't understand this until they are 2 years into building the thing.

LLMs have their merits and he sometimes aludes to them, although it almost feels accidental.

Also, you don't spend years studying computer science to learn the language/syntax, but rather the concepts and systems, which don't magically disappear with vibe coding.

This whole direction is a cheeky Trojan horse. A dramatic problem, hidden in a flashy solution, to which a fix will be upsold 3 years from now.

I'm excited to come back to this comment in 3 years.

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

#406

Post by diggan »

throwdbaaway wrote: post #342
diggan wrote: post #307

Earlier quoted context omitted.

Which model? Just tried a bunch of ChatGPT, OpenAI's API, Claude, Anthropic's API and DeepSeek's API with both chat and reasonee, every single one replied with a single "hi".

o3-mini-2025-01-31 with high reasoning effort replied with "Hi" after 448 reasoning tokens. gpt-4.5-preview-2025-02-27 replied with "Hi!"

> o3-mini-2025-01-31 with high reasoning effort replied with "Hi" after 448 reasoning tokens.

I got "hi", as expected. What is the full system prompt + user message you're using?

https://i.imgur.com/Y923KXB.png

> gpt-4.5-preview-2025-02-27

Same "hi": https://i.imgur.com/VxiIrIy.png

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

#407

Post by diggan »

abdullin wrote: post #299
diggan wrote: post #224

Earlier quoted context omitted.

Well, they're demonstrating it somewhat , it's more of a prototype today. First tell is the low limit, I think the longest task for me been 15 minutes before it gives up. Second tell is still using a chat UI which is simple to implement, easy to implement and familiar, but also kind of lazy. There should be a better UX, especially with the new variations they just added. From the top of my head, some graph-like UX mi…

I guess, it depends on the case and the approach. It works really nice with the following approach (distilled from experiences reported by multiple companies) (1) Augment codebase with explanatory texts that describe individual modules, interfaces and interactions (something that is needed for the humans anyway) (2) Provide Agent.MD that describes the approach/style/process that the AI agent must take. It should also…

> distilled from experiences reported by multiple companies

Distilled from my experience, I'd still say that the UX is lacking, as sequential chat just isn't the right format. I agree with Karpathy that we haven't found the right way of interacting with these OSes yet.

Even with what you say, variations were implemented in a rush. Once you've iterated with one variation you can not at the same time iterate on another variant, for example.

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

#408

Post by iLoveOncall »

ukprogrammer wrote: post #386
iLoveOncall wrote: post #372

Earlier quoted context omitted.

He lives in a GenAI bubble where everyone is self-congratulating about the usage of LLMs. The reality is that there's not a single critical component anywhere that is built on LLMs. There's absolutely no reliance on models, and ChatGPT being down has absolutely no impact on anything beside teenagers not being able to cheat on their homeworks and LLM wrappers not being able to wrap.

Even an LLM could tell you that that's an unknowable thing, perhaps you should rely on them more.

Has a critical service that you used meaningfully changed to seemingly integrate non-deterministic "intelligence" in the past 3 years in one of its critical paths? I'd bet good money that the answer to literally everyone is no.

My company uses GenAI a lot in a lot of projects. Would it have some impact if all models suddenly stopped working? Sure. But the oncalls wouldn't even get paged.

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

#409

Post by Aeroi »

mrmansano wrote: post #380
Aeroi wrote: post #338

the fanboying for this dudes opinion is insane.

It's pastor preaching for the already converted, not new in the area. The only thing new is that they are selling the kool-aid this time.

It's been a multi-day like conversation where multiple people are trying to obtain the transcripts, publish the text as gospel, and now the video. Like, yes thank you but, holy shit.

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

#410

Post by diggan »

tudorizer wrote: post #405

95% terrible expression of the landscape, 5% neatly dumbed down analogies. English is a terrible language for deterministic outcomes in complex/complicated systems. Vibe coders won't understand this until they are 2 years into building the thing. LLMs have their merits and he sometimes aludes to them, although it almost feels accidental. Also, you don't spend years studying computer science to learn the language/synt…

> English is a terrible language for deterministic outcomes in complex/complicated systems

I think that you seem to be under the impression that Karpathy somehow alluded to or hinted at that in his talk, which indicates you haven't actually watched the talk, which makes your first point kind of weird.

I feel like one of the stronger points he made, was that you cannot treat the LLMs as something they're explicitly not, so why would anyone expect deterministic outcomes from them?

He's making the case for coding with LLMs, not letting the LLMs go by themselves writing code ("vibe coding"), and understanding how they work before attempting to do so.

Post reply on HN