Live data from Hacker News

Genie 2: A large-scale foundation world model

deepmind.google

81–90 of 436 posts

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

#81
post #41

Earlier quoted context omitted.

> the squealing carcass called Gemini Have you used Gemini? It seems every bit as good as ChatGPT.

It is not nearly as good. I tried the free trial and cancelled it before it was over.

https://www.cnet.com/tech/services-and-software/chatgpt-vs-g...

https://www.tomsguide.com/ai/google-gemini-vs-openai-chatgpt

It won these shootouts and that's been my experience also, when I need to use AI (extremely rare) I just use the Google Gemini free one. I feel like this is how most people will use AI and why it is doomed to be the ultra low margin grocery store business instead of the huge cash cow business people think it will be.

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

#82

I can.. see this being used to solve crime, even solving unsolved mysteries and cold cases, among other alternative applications.

I don't understand your line of reasoning here. Are you picturing a situation where you would take a photo of a crime scene, and "jump" into a virtual model created from the photo, to help generate intuitions about where to go look for clues ? Kinda like the CSI "enhance quality" meme, but on steroids ?

That would be fun to use, but ultimately pointless. An AI model will generate things that are _statistically plausible_ ; solving crimes usually requires finding out the _truth_.

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

#84
post #4

This is.. super impressive. I'd like to know how large this model is. I note that the first thing they have it do is talk to agents who can control the world gen; geez - even robots get to play video games while we work. That said; I cannot find any: - architecture explanation - code - technical details - API access information Feels very DeepMind / 2015, and that's a bummer. I think the point of the "we have no moat…

[flagged]

I'm guessing the compute needed to create this particular demo is prohibitive for any real-world use case, but it's an impressive proof of concept. Someday it will run on a phone tho.

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

#85
post #13

Related recent project you can try out yourself (Chrome only) which hallucinates new frames of a Minecraft style game: https://oasis.decart.ai/ That one would reimagine the world any time you look at the sky or ground. Sounds like Genie2 solves that: "Genie 2 is capable of remembering parts of the world that are no longer in view and then rendering them accurately when they become observable again."

There is another recent project that is more general game generation very similar to Genie 2. I can't remember the name.

GameGen-X came out last month. https://arxiv.org/html/2411.00769v1

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

#86
post #13

Related recent project you can try out yourself (Chrome only) which hallucinates new frames of a Minecraft style game: https://oasis.decart.ai/ That one would reimagine the world any time you look at the sky or ground. Sounds like Genie2 solves that: "Genie 2 is capable of remembering parts of the world that are no longer in view and then rendering them accurately when they become observable again."

RE: "Genie 2 is capable of remembering parts of the world that are no longer in view and then rendering them accurately when they become observable again." -- This claim is almost certainly wildly misleading. This claim is technically true if there's any scenario where their agent, eg, briefly looked down at the ground and then back up at the sky and at least one of the clouds in the sky was the same as before looking down. However, I expect most people will interpret the claim far more broadly than the model can support. It's classic weasel wording.

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

#90
post #41

Earlier quoted context omitted.

[flagged]

> the squealing carcass called Gemini Have you used Gemini? It seems every bit as good as ChatGPT.

I benchmark these for my job.

Just did one a couple days ago, fortitously.

Gemini Advanced at $20/month is the worst of any commercial model. One constant over the last 6 months is it is indistinguishable from Llama 3.1 8B with search snippets.

Post reply on HN