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

Karpathy’s Pelican

twitter.com

Post reply on HN
Print view View original

131–140 of 461 posts

Re: Karpathy’s Pelican

#131

Post by khazhoux »

throwaway27448 wrote: post #9
misiti3780 wrote: post #5

Earlier quoted context omitted.

i remember this failing, but this product looks pretty useful in 2026

It looked pretty useful in 2009. Failing to push it was baffling then too.

This was the first big secret “you’re not allowed to know what this is or talk to anyone about it” project inside Google. They wanted to be left alone and especially to not have to integrate with the mail team. And everyone heard rumors of gigantic bonuses if they hit whatever milestones, which wasn’t a thing for other projects. All this made them isolated within the company, and when the project was an obvious flop, you didn’t see anyone rushing over to help them.

Re: Karpathy’s Pelican

#132

Post by ben_w »

revel wrote: post #61
matsemann wrote: post #41

I'm pretty tired of the "Y made this game in Z tokens" all over the internet last week. They look impressive, and it's cool that it's even possible, but they're useless as games. None of them are any fun. They're like the most boring variant of basic controllers you can imagine. None have any cool mechanics. None have any tweaks made from hours and hours of testing. All have the same cel-shader.

LLMs are bad at creative work and I don’t see them improving any time soon. Try asking an agent to write a story about raccoons. It will almost certainly involve either stealing food or raiding trash with a 50% chance of having a character named Pip. If a location is mentioned, it’ll be Elm Street. I guess this is the average story and, similarly, the average game is boring and predictable

While they are bad at creative work*, this reasoning isn't going to show it.

If you took the best, most creative, human writer in the world, and for thought experiment reasons they had amnesia (to mimic AI blank context windows) specifically while you asked them for a story idea 100 times in a row, my expectation is that this human would also give you the same idea at least 80 times out of that 100.

* still better than the mean human, but even the top 0.1% of humans aren't all professional authors.

Re: Karpathy’s Pelican

#133

Post by morbicer »

barrenko wrote: post #124

This has started to feel a bit like the beginning of railroads and then the steampunk fiction of "let's just build railroads to everywhere". We don't need it and there's no use for it. As with painting, after a while there's nothing really new to paint, we genuinely need 0 new software. We need to fix our broken physical world, our social lives, our kids and what's left of our democracies. This software crap is done,…

Beautifully put. Software is ok. It's not going to fix our broken world. If AI could at least find cure for some illnesses that would be great. Or decrease social injustice and help with global warming. But it's likely going to do the exact opposite.

Re: Karpathy’s Pelican

#134

Post by kiwibyproxy »

dundarious wrote: post #51

I can forgive the modeling being godawful jank (windows floating in the air, disconnected from the house). But I expected it to have a better understanding of the text. Instead, we have Bilbo's "disappearance" interpreted as him magically transporting or cloaking, and similarly for his reappearance.

which to be fair, happens just a few paragraphs later :) I first watched without sound and thought "oh that's the birthday speech disappearance"

Re: Karpathy’s Pelican

#135

Post by dofm »

jmugan wrote: post #93

A lot of people are posting here about how bad the end product is, but that is kind of the point. Models have moved beyond generating images to a new kind of benchmark that better exposes understanding of the physical world, and we can use benchmarks like this to measure future progress. (Of course, it will have to be a qualitative/subjective measurement.)

But it's another benchmark on how good models are at generating intensely average, unwanted things with unthinking design. Just scaled up.

Re: Karpathy’s Pelican

#137

Post by baron816 »

billyp-rva wrote: post #58
baron816 wrote: post #11

IMO, the area where AI is going to be most useful over the next couple years is in developing manufacturing processes top to bottom. Maybe a million token budget is too small, but something like "design me a sneaker and all the equipment to manufacture it autonomously".

> "design me a sneaker and all the equipment to manufacture it autonomously" We already have sneaker designs and the equipment to manufacture them. Whatever it spits out is going to be, at best, a mediocre clone of something that already exists. What exactly is the point?

Sneaker manufacturing is a very manual process. Nike famously tried to automate it and failed (https://www.wsj.com/economy/trade/why-its-so-difficult-for-r...). Designing and constantly changing all the processes is more expensive than just having people do it by hand. If you can have AI design the process, then the equation switches.

Re: Karpathy’s Pelican

#138

Post by emp17344 »

ben_w wrote: post #132
revel wrote: post #61

Earlier quoted context omitted.

LLMs are bad at creative work and I don’t see them improving any time soon. Try asking an agent to write a story about raccoons. It will almost certainly involve either stealing food or raiding trash with a 50% chance of having a character named Pip. If a location is mentioned, it’ll be Elm Street. I guess this is the average story and, similarly, the average game is boring and predictable

While they are bad at creative work*, this reasoning isn't going to show it. If you took the best, most creative, human writer in the world, and for thought experiment reasons they had amnesia (to mimic AI blank context windows) specifically while you asked them for a story idea 100 times in a row, my expectation is that this human would also give you the same idea at least 80 times out of that 100. * still better th…

Better than the mean human? The mean human creates far better stories while daydreaming. AI enthusiasts have such a distorted view of human capability, it’s bizarre.

Re: Karpathy’s Pelican

#139

Post by MasterScrat »

revel wrote: post #61
matsemann wrote: post #41

I'm pretty tired of the "Y made this game in Z tokens" all over the internet last week. They look impressive, and it's cool that it's even possible, but they're useless as games. None of them are any fun. They're like the most boring variant of basic controllers you can imagine. None have any cool mechanics. None have any tweaks made from hours and hours of testing. All have the same cel-shader.

LLMs are bad at creative work and I don’t see them improving any time soon. Try asking an agent to write a story about raccoons. It will almost certainly involve either stealing food or raiding trash with a 50% chance of having a character named Pip. If a location is mentioned, it’ll be Elm Street. I guess this is the average story and, similarly, the average game is boring and predictable

I don't think this is bad. I want tools that deliver what I envision. The fact that given the same input, I get similar output is something I would rather consider a positive. The imagination should come from the human at the wheel.

Re: Karpathy’s Pelican

#140

Post by knollimar »

dofm wrote: post #128
eterm wrote: post #119

Earlier quoted context omitted.

Right, but the new phenomena is that we've lost a signal. Used to be if a game looked that good it probably had time spent on the game part too.

And you could almost always tell a "game construction kit" game from a real one, couldn't you? Whatever was missing from those is always missing from these, and it seems just as rare that the game will reach much beyond its origins. Whenever I think about AI games I find myself thinking about Tiny Wings, Flappy Bird and Angry Birds. Three simple, elegant games. It is easy to see what makes Tiny Wings so completely lo…

it's crazy because I've seen "implement flappy bird on an FPGA with VGA output" as a 2 week project to a human as something immediately understandable.

Maybe it's the smoothness of motion that makes these games understandable and LLMs seem to consistently fail at that. Ask them to do something snowboarding and they go really hard on the physics since it seems like they don't know what kind of approxmations feel good.

Post reply on HN