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

711–720 of 827 posts

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

#711

Post by diggan »

mupuff1234 wrote: post #416
Aeroi wrote: post #338

the fanboying for this dudes opinion is insane.

Yeah, not sure I ever saw anything similar on HN before, feels very odd. I mean the talk is fine and all but that's about it?

> Yeah, not sure I ever saw anything similar on HN before, feels very odd.

What exactly have you been seeing here on HN? I've been reading through most of the comments in this submission, since it was submitted yesterday, and none of it seems to be "fanboying" (maybe I misunderstand the term?) but discussions about where LLMs fit in the software development workflow.

Some people find some parts interesting, others obvious, others think he's selling something, others find the analogies lacking, but I've seen no "fanboy" comments like what parent seemed to exclusively see here.

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

#712

Post by diggan »

guappa wrote: post #180
shakna wrote: post #167

Earlier quoted context omitted.

Probably too much snark from me. But the gulf between interpreter and compiler can be decades of work, often discovering new mathematical principles along the way. The idea that you're fine to risk everything, in the way agentic things allow [0], and want that messing around with raw memory is... A return to DOS' crashes, but with HAL along for the ride. [0] https://msrc.microsoft.com/update-guide/vulnerability/CVE-2…

Ah don't worry, llms are a return to crashes as it is :) The other day it managed to produce code that made python segfault.

> produce code that made python segfault

To be fair, that's pretty easy for a human to do too.

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

#713

Post by diggan »

yahoozoo wrote: post #208

I was trying to do some reverse engineering with Claude using an MCP server I wrote for a game trainer program that supports Python scripts. The context window gets filled up _so_ fast. I think my server is returning too many addresses (hex) when Claude searches for values in memory, but it’s annoying. These things are so flaky.

Yeah, usually I'd steer my agents to never use the output directly from any command, and instead redirect it to a logfile, then force it to search/grep stuff directly from the log-file instead of just getting all the outputs at all times. Seems to work OK.

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

#714

Post by diggan »

manyaoman wrote: post #699
fnord77 wrote: post #47

Him claiming govts don't use AI or are behind the curve is not accurate. Modern military drones are very much AI agents

Governments obviously lead in military tech, but do you think they have access to better AI (in general) than consumers? Unless they do, I think it's fair to say that governments are behind the curve, since consumers tend to adopt things more quickly.

> but do you think they have access to better AI (in general) than consumers?

Absolutely. One of the top AI labs today is OpenAI, with ties to the US military, not least through Paul M. Nakasone, but also active contracts with the military, announced just a couple of days ago

> In June 2025, the U.S. Department of Defense awarded OpenAI a $200 million one-year contract to develop AI tools for military and national security applications. OpenAI announced a new program, OpenAI for Government, to give federal, state, and local governments access to its models, including ChatGPT. - https://en.wikipedia.org/wiki/OpenAI#Use_by_military

It would be foolish to assume those collaborations are just about API usage with the same models that consumer have access to, there is definitely deeper collaborations than that.

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

#715

Post by diggan »

sensanaty wrote: post #552
diggan wrote: post #218

Earlier quoted context omitted.

Microsoft (famously developing somewhat popular office-like software) seems to be going in the direction of almost forcing developers to use LLMs to assist with coding, at least going by what people are willing to admit publicly and seeing some GitHub activity. Google (made a small browser or something) also develops their own models, I don't think it's far fetched to imagine there is at least one developer on the Ch…

> Microsoft https://news.ycombinator.com/item?id=44050152 Very impressive indeed, not a single line of any quality to be found despite them forcing it on people.

Lets not change the goalpost, parent asked for any examples of software written with LLMs, and regardless if the output is quality or not, that is one example. Besides, Microsoft isn't really known for their high code quality, so I'm not even sure using even dumb LLMs/tools like Copilot would actually have a negative effect.

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

#716

Post by diggan »

cobertos wrote: post #467
oc1 wrote: post #440

Earlier quoted context omitted.

AI is all about context window. If you figured out the context problem, you will see that all these "AI is bullshit, it doesn't work and can't produce working code" goes away. Same for everything else.

Untrue. I find problems with niche knowledge, heavy math, and/or lack of good online resources to be troublesome for AI. Examples so far I've found of consistent struggle points are shaders, parsers, and streams (in Nodejs at least) Context window will solve a class of problems, but will not solve all problems with AI.

I think probably the biggest help I've got from LLMs is things that are "niche" knowledge, for me. Things like "I need a function heavy in math that when I give X and Y, it returns Z" I could have struggled with for days sometimes when I'm writing games for fun, but with LLMs I can have it done and move on in a couple of minutes, the most time consuming part is writing the tests and overall testing, but I no longer spend days just trying to understand enough math to actually write the thing.

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

#717

Post by diggan »

andrethegiant wrote: post #61
practal wrote: post #42

Earlier quoted context omitted.

If you have different representations of the same thing (llms.txt / HTML), how do you know it is actually equivalent to each other? I am wondering if there are scenarios where webpage publishers would be interested in gaming this.

is a standards-friendly way to semantically represent the same content in a different format

Also HTTP headers Accept/Content-Type which in theory could let you serve HTML, XML and JSON all under the same URL/URI but depending on Accept values.

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

#718

Post by diggan »

johnwheeler wrote: post #651

https://github.com/screencam/typescript-mcp-server I've been working on this project. I built this in about two days, using it to build itself at the tail end of the effort. It's not perfect, but I see the promise in it. It stops the thrashing the LLMs can do when they're looking for types or trying to resolve anything like that.

> Traditional: Read 5000 lines → Find method → Replace → Write 5000 lines

What of today's agents work like this? None of the ones I've tried would do something like that, but instead would grep/search the file, then do a smaller edit (different tools do those in different ways).

Overall, it does feel like a strawman argument against "Traditional" when almost none of the tooling actually works like that.

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

#719

Post by diggan »

politelemon wrote: post #120

The beginning was painful to watch as is the cheering in this comment section. The 1.0, 2.0, and 3.0 simply aren't making sense. They imply a kind of a succession and replacement and demonstrate a lack of how programming works. It sounds as marketing oriented as "Web 3.0" that has been born inside an echo chamber. And yet halfway through, the need for determinism/validation is now being reinvented. The analogies make…

> The beginning was painful to watch as is the cheering in this comment section.

Yours is the second comment claiming there is "cheering" and "fanboying" in this comment section. What comments are you talking about? I've read through this submission multiple times since yesterday, yet I've seen none of that. What specific comments are the "cheering" ones?

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

#720

Post by bgwalter »

aaron695 wrote: post #686
OJFord wrote: post #661

I'm not sure about the 1.0/2.0/3.0 classification, but it did lead me to think about LLMs as a programming paradigm: we've had imperative & declarative, procedural & functional languages, maybe we'll come to view deterministic vs. probabilistic (LLMs) similarly. def __main__: You are a calculator. Given an input expression, you compute the result and print it to stdout, exiting 0. Should you be unable to do this, you…

[flagged]

> It makes no sense at all, it's cuckooland, are you all on crazy pills?

Frequent LLM usage impairs thinking. The LLM has no connection to reality, and it takes over people's minds.

Post reply on HN