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

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

dwarkesh.com

Post reply on HN
Print view View original

821–830 of 1001 posts

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

#821

Post by weatherlite »

sarchertech wrote: post #761
sothatsit wrote: post #699

Earlier quoted context omitted.

I don't think he is saying agents are not useful at all, just that they are not anywhere near the capability of human software developers. Karpathy later says he used agents to write the Rust translation of algorithms he wrote in Python. He also explicitly says that agents can be useful for writing boilerplate or for code that can be very commonly found online. So I don't think he is saying they are not useful at all…

I’m not saying he’s saying agents aren’t useful at all. It’s literally in the quotes I provided that he says they are useful for some subset of tasks. I’m saying that he is answering the question “are agents useful at all”. not “can agents replace humans”. His answer is mostly not. He generally prefers autocomplete. But they are useful for some limited tasks.

> I’m not saying he’s saying agents aren’t useful at all

I'm not saying you're saying he's saying agents aren't useful at all

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

#822

Post by thadk »

getnormality wrote: post #804

One of the most brilliant AI minds on the planet, and he's focused on education. How to make all the innovation of the last decade accessible so the next generation can build what we don't know how to do today. No magical thinking here. No empty blather about how AI is going to make us obsolete with the details all handwaved away. Karpathy sees that, for now, better humans are the only way forward. Also, speculation…

Imagine a PhD mortally terrified of exceptions!

Now I see why Karpathy was talking of RL up-weights as if they were a destructive straw-drawn line of a drug for an LLM's training.

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

#823

Post by subjectivationx »

rhetocj23 wrote: post #394
qlm wrote: post #292

Earlier quoted context omitted.

Language isn't thought. It's a representation of thought.

Its very interesting to see how many people struggle to understand this.

We are paying the price now for not teaching language philosophy as a core educational requirement.

Most people have had no exposure to even the most basic ideas of language philosophy.

The idea all these people go to school for years and don't even have to take a 1 semester class on the main philosophical ideas of the 20th century is insane.

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

#824

Post by dahart »

stouset wrote: post #652
ziofill wrote: post #545

Earlier quoted context omitted.

I agree with this. A metaphor I like is that the reason why humans say the night sky is beautiful is because they see that it is, whereas an LLM says it because it’s been said enough times in its training data.

To play devil’s advocate, you have never seen the night sky. Photoreceptors in your eye have been excited in the presence of photons. Those photoreceptors have relayed this information across a nerve to neurons in your brain which receive this encoded information and splay it out to an array of other neurons. Each cell in this chain can rightfully claim to be a living organism in and of itself. “You” haven’t directly…

This comment illustrates the core problem with reductionism, a problem that has been known for many centuries, that “a system is composed entirely of its parts, but the system will have features that none of the parts have” [1] thus fails to explain those features.

The ‘you have never seen’ assertion feels like a semantic ruse rather than a helpful observation. So how do you define “you” and “see”? If I accept your argument, then you’ve only un-defined those words, and not provided a meaningful or thoughtful alternative to the experience we all have and therefore know exists.

I have seen the night sky. I am made of cells, and I can see. My cells individually can’t see, and whether or not they can claim to be individuals, they won’t survive or perform their function without me, i.e., the rest of my cells, arranged in a very particular way.

Today’s AI is also a ruse. It’s a mirror and not a living thing. It looks like a living thing from the outside, but it’s only a reflection of us, an incomplete one, and unlike living things it cannot survive on its own, can’t eat or sleep or dream or poop or fight or mate & reproduce. Never had its own thoughts, it only borrowed mine and yours. Most LLMs can’t remember yesterday and don’t learn. Nobody who’s serious or knows how they work is arguing they’re conscious, at least not the people who don’t stand to make a lot of money selling you magical chat bots.

[1] https://en.wikipedia.org/wiki/Reductionism#Definitions

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

#825

Post by newsclues »

jaza wrote: post #636
newsclues wrote: post #278

Earlier quoted context omitted.

What was the last example where humans succeeded at a hard problem like that? Space flight?

We've "succeeded" at space flight about as much as we've "succeeded" at AI. Yay, man on the moon! Over half a century later, and it turns out that the "next small step" - man on Mars - isn't so small and still hasn't been achieved. Anything remotely resembling sci-fi-style ubiquitous space travel remains exactly that - sci-fi!

Flying a plane and intercontinental flight are different levels of the same remarkable achievement.

A man on the moon, or the SpaceX rockets that land and can rapidly relaunch, both feel like hard problems that have been solved, although it’s not the next hard step of intergalactic space travel.

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

#826

Post by modeless »

chronci739 wrote: post #691
modeless wrote: post #648

Earlier quoted context omitted.

Yeah, I see people pooh-poohing the idea of humanoid robots being useful this decade, saying it will take at least 20 years. Oh yeah? Instead of 5 years to render all human labor obsolete, it will take 20? The magnitude of that change is so large that the implications of it happening anytime in our lifetimes are too big to ignore. The important thing is that this is not going to be perpetually 20 years in the future…

> This is something that will happen. Not in our lifetime. The iPhone came out less than 20 years ago. And what, you scan QR codes at restaurants with iphones?

If robots "only" have an iPhone-sized impact on the world I think that would be surprising but also still a huge deal worth caring about.

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

#827

Post by Rover222 »

musebox35 wrote: post #820
Rover222 wrote: post #782

Earlier quoted context omitted.

Both Karpathy and must have explained it many times. The additional sensors add more signal than noise in the end. You also then have to decide which sensor system is correct any time they disagree. Also the entire road system is designed for vision. Lidar cannot read signs, see colors, etc. Humans can drive with two eyes. It's not insane to think computers can do it with 7 or 8 cameras. As someone who has used Tesla…

And that last 5% is the toughest nut to crack. There is a reason waymo is way ahead even if they can not scale. Cameras are passive devices with relatively poor dynamic range and low light behavior. They are nowhere near a match/replacement for the human eye. Just try to picture a 5 year old at dusk or indoors and what you see will not be what you get.

Agree that the last fiew percentage points are exponentially more difficult each step of the way. What's your metric for saying Waymo is ahead, in terms of tech? They are strictly geo fenced, limited to specific road types, and often get stuck/confused. Also their system is very expensive, and not scalable to million of cars. Your point about cameras seems odd. Cameras have much better low light performance than human eyes. And cars have headlights.

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

#828

Post by jobs_throwaway »

whiplash451 wrote: post #414
hax0ron3 wrote: post #355

If the transcript is accurate, Karpathy does not actually ever, in this interview, say that AGI is a decade away, or make any concrete claims about how far away AGI is. Patel's title is misleading.

Did the same with Sutton (LLMs are a dead end) when Sutton never said this in the conversation.

He didn't say those words exactly but he did say

"The scalable method is you learn from experience. You try things, you see what works. No one has to tell you. First of all, you have a goal. Without a goal, there’s no sense of right or wrong or better or worse. Large language models are trying to get by without having a goal or a sense of better or worse. That’s just exactly starting in the wrong place."

and a bunch of similar things implying LLMs have no hope of reaching AGI

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

#829

Post by rhetocj23 »

aik wrote: post #803
Culonavirus wrote: post #554

Earlier quoted context omitted.

Oh look, people with skin in the AI game insist AI is not a massive bubble. More news at 11.

We’re a regular old SaaS company that has figured out how to add massive value using AI. I am making no statements about valuations and bubbles. I’m actually guessing there is some bubble / overhype. That doesn’t mean it isn’t still incredibly valuable.

Link? And explain in detail, incrementally, what you have done so we can analyse it for ourselves?

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

#830

Post by TZubiri »

getnormality wrote: post #804

One of the most brilliant AI minds on the planet, and he's focused on education. How to make all the innovation of the last decade accessible so the next generation can build what we don't know how to do today. No magical thinking here. No empty blather about how AI is going to make us obsolete with the details all handwaved away. Karpathy sees that, for now, better humans are the only way forward. Also, speculation…

>An exception is a signal of uncertainty indicating that you need to learn more about your problem

No, that would be a warning. Ab exceprion is a signal something failed and it was impossible to continue

Post reply on HN