Andrej Karpathy – It will take a decade to work through the issues with agents

Post by JimDabell » Fri, Oct 17, 2025, 6:46 PM UTC

sdenton4 wrote: post #83

Imnimo wrote: post #68

>What takes the long amount of time and the way to think about it is that it’s a march of nines. Every single nine is a constant amount of work. Every single nine is the same amount of work. When you get a demo and something works 90% of the time, that’s just the first nine. Then you need the second nine, a third nine, a fourth nine, a fifth nine. While I was at Tesla for five years or so, we went through maybe three…

Ha, I often speak of doing the first 90% of the work, and then moving on to the following 90% of the work...

Post by AnimalMuppet » Fri, Oct 17, 2025, 6:49 PM UTC

throw54465665 wrote: post #150

zeroonetwothree wrote: post #117

Earlier quoted context omitted.

I wouldn’t say that any specific skill (like literacy) is required to have intelligence. It’s more the capability to learn skills and build a model of the world and the people in it using abstract reasoning. Otherwise we would have to say that pre-literacy societies lacked intelligence, which would be silly since they are the ones that invented writing in the first place!

[flagged]

Post by rwj » Fri, Oct 17, 2025, 6:50 PM UTC

sysguest wrote: post #159

ekjhgkejhgk wrote: post #134

Earlier quoted context omitted.

The interview which I've watched recently with Rich Sutton left me with the impression that AGI is not just a matter of adding more 9s. The interviewer had an idea that he took for granted: that to understand language you have to have a model of the world. LLMs seem to udnerstand language therefore they've trained a model of the world. Sutton rejected the premise immediately. He might be right in being skeptical here…

yeah that "model of the world" would mean: babies are already born with "the model of the world" but a lot of experiments on babies/young kids tell otherwise

Post by adastra22 » Fri, Oct 17, 2025, 6:51 PM UTC

simpleladle wrote: post #60

adastra22 wrote: post #24

Earlier quoted context omitted.

Machine learning as a descriptive phrase has stopped being relevant. It implies the discovery of information in a training set. The pre-training of an LLM is most definitely machine learning. But what people are excited and interested in is the use of this learned data in generative AI. “Machine learning” doesn’t capture that aspect.

But the things we try to make LLMs do post-pre-training are primarily achieved via reinforcement learning. Isn't reinforcement learning machine learning? Correct me if I'm misconstruing what you're trying to say here

Post by gtirloni » Fri, Oct 17, 2025, 6:54 PM UTC

tasuki wrote: post #166

TheBlight wrote: post #149

I knew this once I heard OpenAI was going to get into the porn bot business. If you have AGI you don't need porn bots.

Why not?