Cool demos, but does anybody else find it troubling that companies keep claiming this tech can train robots? It’s so obviously not representative of the real world.

▲

overfeed 5 hours ago | parent | next [-]

> Cool demos, but does anybody else find it troubling that companies keep claiming this tech can train robots? It’s so obviously not representative of the real world.

A proprietary version of this technology is already training Waymo's robot cars. Waymo blogged about their DeepMind-provided world model a year ago.

> It’s so obviously not representative of the real world.

That entirely depends on the use-case and level of granularity needed, doesn't it?

▲

sosodev 4 hours ago | parent | next [-]

Isn’t the Waymo version used for some small portion of the dataset to cover edge cases that they have no real data for?

This blog is suggesting that we could train robots using only a world model.

> Do you have a basis for this idea, or it's a gut-feel?

My basis is that I’m a real human and these videos look far more like bad video games than the real life I’ve experienced.

	▲	overfeed 2 hours ago \| parent [-]
		Apologies for the incidindiary wording of my initial concluding question (which I immediately edited to be less flamey, and hopefully more cogent). > Isn’t the Waymo version used for some small portion of the dataset to cover edge cases that they have no real data for? The usecases in the blog post[1] paint a much more expansive picture of how Wayne used Genie 3: including enabling transfer learning to convert single-camera dashcam footage into simulated full-take sensor data (multi-camera and Lidar), which is extremely valuable. > My basis is that I’m a real human and these videos look far more like bad video games than the real life I’ve experienced 1:1 fidelity isn't required for a lot of training tasks; for instance, computer vision can do a lot with 32x32 resolution images, which is trash quality to humans. I think the potential customers for this technology will constrain the model to fit their use cases. 1. https://waymo.com/blog/2026/02/the-waymo-world-model-a-new-f...

▲

whywhywhywhy 4 hours ago | parent | prev [-]

Train robots feels like the academia justification excuse when we really know the value of this will be entertainment.

▲

ActionHank 5 hours ago | parent | prev | next [-]

Elon revealed to the world that you can just lie and people will give you money, so now they all lie.

	▲	solyin an hour ago \| parent [-]
		Elizabeth shows us that it has to be the right lie though.

▲

TomGarden 5 hours ago | parent | prev | next [-]

This stands out to me too, it seems to be representative of physics the way a video game is representative of physics, likely less so given the non-determinism

▲

killerstorm 2 hours ago | parent | prev | next [-]

Gemini Robotics is a thing: https://deepmind.google/models/gemini-robotics/

They clearly know how to train robots. Unless Genie is a misdirection...

▲

arjie 4 hours ago | parent | prev | next [-]

Subreddits where SolidGoldMagikarp was counting to a million in comments were used to train a machine that then writes web pages that more machines are trained on. The sheer volume of data with sufficient verisimilitude generated by these models may well work to train robots. It might not, but I think it might well work. The idea that things have to be high-fidelity high-quality to yield good results has clearly not held up in other domains.

▲

jayd16 3 hours ago | parent | prev | next [-]

I guess they want military contracts or something.

▲

kragen 2 hours ago | parent | prev [-]

You can get useful levels of robot training out of Super Mario Bros 2; see John Carmack's talk from last year: https://www.youtube.com/watch?v=4epAfU1FCuQ

When your training simulation fails to reproduce certain features of the real-world environment, your control model is likely to have trouble with them, so more realistic simulations are helpful, but even fairly unrealistic training simulations are pretty useful.

In general, training neural networks to control robots is going to require simulations, ideally differentiable simulations. Even if they aren't differentiable, though, you can SGD your simulated training runs across a data center in a way that you just can't with training runs on Roombas.