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

Genie 2: A large-scale foundation world model

deepmind.google

Post reply on HN
Print view View original

221–230 of 436 posts

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

#221

Post by luckydata »

margalabargala wrote: post #108
xnx wrote: post #41

Earlier quoted context omitted.

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

Something Gemini utterly fails at compared to ChatGPT and Claude is its willingness to contradict you if you are wrong and it is correct. Ask any of the three models basic facts and they'll answer correctly. If you tell ChatGPT or Claude Sonnet they are incorrect, they will push back, if they really think they are right. Gemini will fall over itself to agree with you that it's terribly sorry, you are right, the sky i…

This made me laugh out loud so much because while it's not as true anymore, it's a pretty good distillation of how unwilling try he average Googler is to be disagreeable and I guess Gemini absorbed some of that from the people that worked on it. Just like normal software AIs seem to be the expression of the organization that produces it but in this case it's easier to spot it as it gives it a sort of "persona".

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

#222

Post by lxgr »

xnx wrote: post #41
justlikereddit wrote: post #30

Earlier quoted context omitted.

[flagged]

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

I have, and it's terrible in exactly the way GP describes it.

It won't talk to me about anything involving the word "president" or anything related to the US political system, even very procedural/hopefully uncontroversial questions such as "who appoints , and is the appointment confirmed in congress or not".

That's only one example; it generally refuses so many things (and often even lies about "not being able to", despite sometimes leaking the correct answer for a second and then overwriting that with the lie) that I've given up on it – for the second time.

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

#223

Post by wongarsu »

cptroot wrote: post #105

For all that this is lauded as a "prototyping tool", it's frustrating to see Genie2 discarding entire portions of the concept art demo. The original images drawn by Max Cant have these beautiful alien creatures. Large ones floating, and small ones being herded(?). Genie2 just ignores these beautiful details entirely: > That large alien? That's a tree. > That other large alien? It's a bush. > That herd of small creatu…

Yeah, those two demos fell flat for me. The model performing badly on inputs far outside the training data is fine, but those two videos belong in the outtakes section or maybe a limitations section, not next to text lauding the "out-of-distribution generalization capabilities". The videos show the opposite of what's claimed.

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

#224

Post by luckydata »

int_19h wrote: post #203
talldayo wrote: post #73

Earlier quoted context omitted.

I honestly disagree. Google is the only FAANG member that has made AI features I actually consider helpful. Furthermore, I think their strategy is heads-and-shoulders above Microsoft and Apple, who are both desperately paying other businesses to do the "AI stuff" for them. Google is at least doing it in-house, and I don't feel at all like they're behind OpenAI's integrations.

Apple use of AI is based on their own models first, and they have spent an inordinate amount of effort on privacy. They fall back onto OpenAI / GPT for stuff that they know their models can't do well, which is not surprising because, like it or not, OpenAI still has the best models on the market - so it makes perfect sense to use them for tough-to-crack stuff (it's certainly better than spectacularly failing like e.g…

If you work from false premises you can get to whatever conclusion you like. Gemini is not bad at all but you do you.

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

#225

Post by talldayo »

int_19h wrote: post #203
talldayo wrote: post #73

Earlier quoted context omitted.

I honestly disagree. Google is the only FAANG member that has made AI features I actually consider helpful. Furthermore, I think their strategy is heads-and-shoulders above Microsoft and Apple, who are both desperately paying other businesses to do the "AI stuff" for them. Google is at least doing it in-house, and I don't feel at all like they're behind OpenAI's integrations.

Apple use of AI is based on their own models first, and they have spent an inordinate amount of effort on privacy. They fall back onto OpenAI / GPT for stuff that they know their models can't do well, which is not surprising because, like it or not, OpenAI still has the best models on the market - so it makes perfect sense to use them for tough-to-crack stuff (it's certainly better than spectacularly failing like e.g…

Privacy is something you give up for all online software. I give up privacy to use YouTube, search engines, Hacker News and even DNS. Unless you can show me the source code for Apple's implementation, I daresay they're forcing you to give up a modicum of privacy too. Same goes for OpenAI's servers, for that matter.

Mostly I dog on Apple and Microsoft because both companies didn't even attempt to compete. They ignored machine learning when it was taking off (FAIR/BERT era), refused to ship complex GPU designs instead of NPUs though it was the only thing working (eg. Nvidia vs everyone else), and they continue to lag behind all their peers in local model development. The most impressive contribution either company makes to Open Source AI is the support they write for Tensorflow and Pytorch - neither of which they even own. If you erased every contribution that Apple and Microsoft made to AI research, the entire field could pick up where it left off tomorrow.

Google's got problems, but AI is something they are far-and-away better at than their contemporaries. It is just Google and Facebook at this echelon, which is embarrassing but frankly expected given how stagnant and anticompetitive Apple and Microsoft's business models have become.

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

#226

Post by sangnoir »

youoy wrote: post #174
sangnoir wrote: post #125

Earlier quoted context omitted.

> Interacting with the actual physical world is very different from creating a video game The major difference being the former scales very poorly for generating training data compared to the latter. Genie 2 is not even a video game and has worse fidelity that video games, the upside is it probably scales even better than video games for generating training scenarios. If you want androids in teal life, Genie 2 (or si…

> The training pipeline will be: raw video -> Genie 2 -> game engine with rules -> physical robot One of those arrows is not like the others

The final step is an oversimplification: purpose-built simulator -> deconstructed robot on a lab workbench -> controlled space -> "real world" with constraints -> real world

Any model would have to succeed in one stage before it can proceed to the next one.

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

#227

Post by miltonlost »

dmarcos wrote: post #217
alphabetting wrote: post #195

Earlier quoted context omitted.

Seems like there's already a lot of slop on steam and I really doubt it will be difficult for quality content to be highlighted even if the amount of games increases 1000x or more

Yeah. Video and Youtube is an example. Filtering is not a hard problem. Mega tons of bad stuff doesn’t bother me.

Love that Youtube filter that spits out what I should consume. Thank you corporate algorithm for telling me what is a good thing to watch

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

#228

Post by wg0 »

Google is not coming slow... This is magic. As a casual gamer and someone wanting to make my own game, this is black magic.

Lighting, gravity, character animation and what not internalized by the model... from a single image...!

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

#229

Post by nopinsight »

The real goal of this research is developing models that match or exceed human understanding of the 3D world -- a key step toward AGI.

A key reason why current Large Multimodal Models (LMMs) still have inferior visual understanding compared to humans is their lack of deep comprehension of the 3D world. Such understanding requires movement, interaction, and feedback from the physical environment. Models that incorporate these elements will likely yield much more capable LMMs.

As a result, we can expect significant improvements in robotics and self-driving cars in the near future.

Simulations + Limited robot data from labs + Algorithms advancement --> Better spatial intelligence

which will lead to a positive feedback loop:

Better spatial intelligence --> Better robots --> More robot deployment --> Better spatial intelligence --> ...

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

#230

Post by onlyrealcuzzo »

wongarsu wrote: post #213
niceice wrote: post #190

Earlier quoted context omitted.

Any estimates of how much one of these cost to generate and keep a minute of context? Secondly, any estimate of how much the price could fall in 5-10 years?

Oasis (the Minecraft world model) can serve about 5 players on 8 H100 in real-time at 20fps in 360p. This is a much more capable model with two orders of magnitude more context. They pretty much say it can't be played real-time, which I read as they generate less than 15fps@240p on 8 GPUs. Probably why they talk so much about using it for AI training and evaluation rather than human use. There is a distilled version…

> Probably why they talk so much about using it for AI training and evaluation rather than human use.

What would they do / how would they use this output to make a better AI?

Post reply on HN