Live data from Hacker News

Genie 2: A large-scale foundation world model

deepmind.google

321–330 of 436 posts

Re: Genie 2: A large-scale foundation world model

#321
post #136

Earlier quoted context omitted.

> Glasses that make the world 20% more pleasant to look at. When AR glasses get good enough to wear all day, I've really been wanting to make a real-life ad blocker.

hallucinogenics are available right now.

blocks more than ads

Re: Genie 2: A large-scale foundation world model

#322

It’s interesting to me that we continue to see such pressure on video and world generation, despite the fact that for years now we’ve gotten games and movies that have beautiful worlds filled with lousy, limited, poorly written stories. Star Wars movies have looked phenomenal for a decade, full of bland stories we’ve all heard a thousand times. Are there any game developers working on infinite story games? I don’t ca…

if stories you're needing, there's an LLM I have to sell you

Re: Genie 2: A large-scale foundation world model

#323
post #318

Earlier quoted context omitted.

That has been the case since art was first industrialized with the printing press. Most of them don’t survive but a significant fraction, if not the vast majority, of books printed in the first century were trashy novels about King Arthur and other fantasies (we know from publisher records and bibliographies that they were very popular but don’t have detailed sales figures to compare against older content like transl…

Sturgeon's law says 90% of everything is crap. If AI pushes that up to 98%, that means you have to look through 5 times as much crap to get the good stuff.

Exactly, “it’s bad now” != “it won’t get worse”.

Re: Genie 2: A large-scale foundation world model

#324

Earlier quoted context omitted.

Counter point: you’d see a corresponding exponential increase in QA labor, and just like with the web, Steam will be absolutely flooded with slop. So I see the most likely outcome is a lot of dogshit and Steam being forced to make draconian moves to protect the integrity of the store.

QAing a game built on a framework where fundamental mechanics are non-deterministic and context-sensitive sounds like a special kind of hell. Not to mention that once you find a bug there's no way to fix it directly, since the source code is an opaque blob of weights, so you just have to RLHF it until it eventually behaves.

And meanwhile you’ve used up .1% of humanity’s remaining carbon budget on each round.

Re: Genie 2: A large-scale foundation world model

#325

Earlier quoted context omitted.

We already knew it's possible from AI minecraft ( https://oasis.decart.ai ). This is just a more impressive version of that, trained on a wider range of games and with more context frames (Oasis has about a second of context, this one a minute). Even the architecture seems to be about the same. Had they released this two months earlier it would have been incredibly impressive. Now it's still cool and inspiring, but n…

This is far beyond Oasis. Oasis had approximately 0 continuity, and the generated world was a blurry mess. This on the other hand actually approaches usability.

And it works on a wide variety of games, instead of just a single one with a relatively consistent art style. On the other hand, Oasis was realtime, while this one is offline; IMO getting the inference speed down was their most impressive feat, as even most decent video gen models are slower than that.

Re: Genie 2: A large-scale foundation world model

#326
post #144

Earlier quoted context omitted.

Not at all. If you look at the publications in the world model space, this can be achieved by any pre-seed company. You just need lots of data, but that's easy to gather in an automated fashion with scripting against game engines. These are basically learning game playthroughs. There are some folks on Discord training a playable Super Smash Bros Melee world model. There's a large corpus of annotated Smash Bros gamepl…

i upvoted you depsite some downvotes because i would like to hear your argument on why this isn't expensive. if it takes a high end consumer video card that costs $1-2k something like 10 seconds to generate a single low res image, presumably effectively real-time video combined with user inputs would be extremely, extremely demanding and therefore expensive. i mean basic napkin math for 25FPS is 250 GPUs, which is ha…

You can run and play several publicly available world models on lower end graphics cards today, and that's with "academic" code: crude, unoptimized, written just to accomplish the science.

Here's one: https://diamond-wm.github.io/

There are lots of others.

From their paper: "Our model was trained in 12 days on an RTX 4090, and can be played at ~10 FPS on an RTX 3090."

This is incredibly accessible, and "this is the worst it will ever look". There are so many low hanging fruit optimizations, and that's without even considering a different model architecture.

Here's someone's world model side project from over two years ago that you can play in your browser:

https://madebyoll.in/posts/game_emulation_via_dnn/demo/

This stuff will be everywhere soon, and you'll be running it on consumer hardware.

Re: Genie 2: A large-scale foundation world model

#327
post #193

Earlier quoted context omitted.

> gamedevs' efficiency requirements These models won't need you to retopo meshes, write custom shaders, or optimize Nanite or Lumen gameplay. They'll generate the final frames, sans traditional graphics processing pipeline. > The 1 thing required for serious developers is control Same with video and image models, and there's tremendous work being done there as we speak. These models will eventually be trained to lear…

> Why do you think that's true? > These will get faster and they will get controllable. Brother, you're preaching to the choir. I've been shilling generative tools for gamedev far harder than you are in your reply. :-) But I'm just relaying to you what actual gamedevs working and writing code right now need and for the foreseeable future for which projects have been started or planned. As Mike Acton says, "the proble…

I've written my own 2D and 3D game engines as well as worked in Unreal. I'm currently working on a controllable diffusion engine using Bevy.

I strongly disagree that you need to cater to existing workflows. There's so much fertile ground in taking a departure. Just look at what's happening with animation and video. People won't be shooting on Arri Alexas and $300,000 glass for much longer.

Re: Genie 2: A large-scale foundation world model

#328

Hey, DeepMind folks, are you listening? Listen. We believe you: you can conquer any virtual world you put your mind to. Minecraft, Starcraft, Warcraft (?), Atari, anything. You can do it! With the power of RL and Neural Nets. Well done. What you haven't been able to do so far, after many years of trying, is to go from the virtual, to the real. Go from Arcanoid to a robot that can play, I dunno, squash, without dying.…

https://sites.research.google/palm-saycan

Tech demo, doesn't generalise.

Re: Genie 2: A large-scale foundation world model

#329

Hey, DeepMind folks, are you listening? Listen. We believe you: you can conquer any virtual world you put your mind to. Minecraft, Starcraft, Warcraft (?), Atari, anything. You can do it! With the power of RL and Neural Nets. Well done. What you haven't been able to do so far, after many years of trying, is to go from the virtual, to the real. Go from Arcanoid to a robot that can play, I dunno, squash, without dying.…

Does Waymo count?

No: remote safety drivers; not DeepMind.

Re: Genie 2: A large-scale foundation world model

#330

It’s interesting to me that we continue to see such pressure on video and world generation, despite the fact that for years now we’ve gotten games and movies that have beautiful worlds filled with lousy, limited, poorly written stories. Star Wars movies have looked phenomenal for a decade, full of bland stories we’ve all heard a thousand times. Are there any game developers working on infinite story games? I don’t ca…

Dwarf fortress is kindof an infinite story generator
Post reply on HN