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

Gemini Robotics 2 brings whole body intelligence to robots

deepmind.google

Post reply on HN
Print view View original

221–230 of 583 posts

Re: Gemini Robotics 2 brings whole body intelligence to robots

#221

Post by SubiculumCode »

heaney-555 wrote: post #78
xgulfie wrote: post #70

Earlier quoted context omitted.

Level 5 automomy has been "only five years away" for 15 years

You're probably thinking of Level 4. No one thought Level 5 was coming before 2050 until a few years ago. And Level 4 is already here, scaling up, while we're seeing the first real signs of Level 5 (Tesla FSD Supervised). The progress here is staggering - I'm not sure why you're so cynical!

For those who live in SF, for example, the progress is obvious. Not speaking for the other person in the thread, but most people have not seen an autonomous car working in real life. Also, many had assumed that freeway driving would happen first (as opposed to complicated city driving), but it turned out that the velocities involved carried too much risk during development. I agree with you, there has been much more progress than I had expected.

Re: Gemini Robotics 2 brings whole body intelligence to robots

#222

Post by ai-x »

snek_case wrote: post #170
ai-x wrote: post #145

Earlier quoted context omitted.

[flagged]

The thing with drug discovery is that the bottleneck is not coming up with new molecules, etc. It's the multi-year, extremely expensive clinical trials that often don't even succeed. I doubt it will be a big moneymaker for Anthropic and OpenAI, because the best they can hope for is to sell a software product to pharma companies, but the pharma companies already have lots of scientists with a million promising molecul…

If you had asked HN 3 years ago what would code-generating market would be, they would have used the exact same reasoning and given us the opinion "Why no one will pay for glorified auto-complete"

Re: Gemini Robotics 2 brings whole body intelligence to robots

#223

Post by yhjc2692 »

YuechenLi wrote: post #152

Running a full LLM to actuate a robot is way too heavy, the minimum latency they can get down to is probably 1-2 seconds even with powerful GPUs, which is not very useful for practical robotics applications. Machine vision should definitely be handled by ML, but motion actuation should be relegated in the realm of traditional PID style linear/nonlinear control. Again, the tech is cool, but the practical usefulness of…

On latency: that is true now but likely false in the near future. There is so much potential left over for optimizing inference further.

On PID: the field has been stuck trying to do analytical/optimization-based control for decades, and end-to-end control has shown incredible performance (e.g., SoTA cost of transport in legged locomotion) and robustness (e.g., not falling over when stepping on a pile of leaves) - while being far scalable (in terms of how fast it is to get a new robot up and running). Which is not to say it's perfect but it seems like it's a step in the right direction.

Better yet, companies like Physical Intelligence are doing good with hierarchical ("fast-slow") architecture to address both the intelligence and latency fronts.

Re: Gemini Robotics 2 brings whole body intelligence to robots

#224

Post by andriy_koval »

ai-x wrote: post #163
yellow_lead wrote: post #150

Earlier quoted context omitted.

OpenAI and Anthropic can't do either of those if they go bankrupt first.

People are going to be surprised by how obscenely profitable both will be in a couple of years

if open models get banned

Re: Gemini Robotics 2 brings whole body intelligence to robots

#225

Post by 0x3f »

ai-x wrote: post #163
yellow_lead wrote: post #150

Earlier quoted context omitted.

OpenAI and Anthropic can't do either of those if they go bankrupt first.

People are going to be surprised by how obscenely profitable both will be in a couple of years

Seems unlikely. All three market leaders have roughly equivalent products, even ignoring the Chinese models. Google is better vertically integrated, as well. They'll end up competing on price, or worst case capacity rationing.

Re: Gemini Robotics 2 brings whole body intelligence to robots

#226

Post by jauntywundrkind »

canyon289 wrote: post #211

I'm a researcher at Deepmind that contributed to these models. (And the opinions here are my own) Just want to say, Deepmind is a great place to work and the only (Edit: one the few unique labs!) lab where you can move from large frontier models (Gemini), frontier open models (Gemma), robotics (what you see here), science (weather, biology, more) and basically any other topic related to intelligence. It's really an i…

Good stuff, yay, good.

I do want to pick a nit in this one,

> and the only lab where

I do think Allen Institution for AI (AI2) is has coverage across most of these domains ( albeit their frontier isn't nearly so far out, hopefully the $152m NSF awarded them + Nvidia is a fruitful partnership there).

For example, robotics: MolmoBot, MolmoSpaces, MolmoAct, https://allenai.org/embodied-ai

Re: Gemini Robotics 2 brings whole body intelligence to robots

#227

Post by xnx »

ai-x wrote: post #163
yellow_lead wrote: post #150

Earlier quoted context omitted.

OpenAI and Anthropic can't do either of those if they go bankrupt first.

People are going to be surprised by how obscenely profitable both will be in a couple of years

Tight competition and "obscenely profitable" rarely go hand-in-hand.

Re: Gemini Robotics 2 brings whole body intelligence to robots

#228

Post by xboxnolifes »

ai-x wrote: post #222
snek_case wrote: post #170

Earlier quoted context omitted.

The thing with drug discovery is that the bottleneck is not coming up with new molecules, etc. It's the multi-year, extremely expensive clinical trials that often don't even succeed. I doubt it will be a big moneymaker for Anthropic and OpenAI, because the best they can hope for is to sell a software product to pharma companies, but the pharma companies already have lots of scientists with a million promising molecul…

If you had asked HN 3 years ago what would code-generating market would be, they would have used the exact same reasoning and given us the opinion "Why no one will pay for glorified auto-complete"

Thr pharma parallel to what tech is doing is replacing clinical trials with AI review. Im not so sure its the same thing.

Re: Gemini Robotics 2 brings whole body intelligence to robots

#229

Post by xnx »

heaney-555 wrote: post #148
xnx wrote: post #15

Earlier quoted context omitted.

Seems smart. There are dozens of companies that make low-margin cars. There's only one company that has managed to make a working autonomous software driver.

Tesla is the only company to have high-margin mass-market EVs.

True, but their margins only exist because of protectionist US laws (i.e. 100% tariff on Chinese competitors).

Re: Gemini Robotics 2 brings whole body intelligence to robots

#230

Post by infinitezest »

luckydata wrote: post #162
kelvinjps10 wrote: post #111

Earlier quoted context omitted.

At that point why not have a human cleaning service it will probably be cheaper

it's not cheap at all and the privacy and logistical implications of letting a stranger inside your house is something not everyone wants to deal with. I would rather have a robot.

Suggesting that a robot is the more privacy friendly option is an interesting take.
Post reply on HN