The Simulator That Could Supercharge Robotics!
❤️ Check out Weights & Biases and sign up for a free demo here:
📝 The paper is available here:
📝 My paper on simulations that look almost like reality is available for free here:
Or this is the orig. Nature Physics link with clickable citations:
🙏 We would like to thank our generous Patreon supporters who make Two Minute Papers possible:
Alex Balfanz, Alex Haro, B Shang, Benji Rabhan, Gaston Ingaramo, Gordon Child, John Le, Juan Benet, Kyle Davis, Loyal Alchemist, Lukas Biewald, Martin, Michael Albrecht, Michael Tedder, Owen Skarpness, Richard Sundvall, Taras Bobrovytsky,, Thomas Krcmar, Tybie Fitzhugh, Ueli Gallizzi.
If you wish to appear here or pick up other perks, click here:
My research:
X/Twitter:
Thumbnail design: Felícia Zsolnai-Fehér –
Robots should be trained in mma gyms. That way we can have robots that have great control over their body and can even join the police force because of that. Also that way we can even recreate that one movie with will smith.
Move along citizen
So much hype! 😂
Just get me a robot who can fold all this damn laundry 😂
https://youtu.be/pycyMUQwiNs?si=l7Rspk-QGJCR5-mI
@@masterodst1 but with AI instead of telepresence.
You don’t need a robot for that
@@blackhole4813 it’s not a need, it’s a want
Can this also train the cobots in restaurant ?
This is very cool for the future of robotics! What a time to be alive!
WHAT A TIME TO BE ALIVE !!!
Can’t simulate down to atom so it will not prove itself to be sustainable and feasible in time.
“Subjective generalization through filtering generated hallucinations of concepts in related subjects” (whatever that will be named in the future) is the best way to train any AI, human brain generalizes the same way, just need more compute and optimization to get there and will eventually do.
so, make AIs imagine? getting the training data would be extremely hard though (since you’d have to do it IRL)
Theres no human on earth who can copy others on an atomic level, but you can still do things and it works, so they dont need to mimic reality 100%
This is called World Model, Smidhuber made a paper about that with a student.
4:45 that joke caught me off guard lmao. Love your content keep it up
What a time to be alive and witness all this advancements.
What a time to still have permission from our AI masters to be alive!
Be careful, cats will invent a robot to pet and feed them and they’ll start replacing us.
We’ll be obsolete and cats won’t need us anymore!
AI is dangerous indeed, but it’s the cats we should be worried about 😂
Absolutely incredible. So important process innovation. Thanks for sharing this.
I would just like to point out that I, according to all relevant regional legal requirements, have proven myself to be a safe driver consistently and in a multitude of simulated scenarios.
I also still mount the kerb on occasion, and my car’s exterior is home to more than a handful of “not quite insignificant” dents and scratches.
Make of that information what you will…
Again with this “robots in simulations” topic??
I’m sure these robots will be used for folding laundry and nothing destructive…
Pattycake always ends in a brutal fight.
Mightily!
With all the accidents I see on automated vehicles a normal driver could easily interpret I believe we need a connected system all around with multiple sensors, just one type is not sufficient.
The Increasingly Inaccurately Named Two Minute Papers channel gives Douglas Adams vibes.
Out of an abundance of caution, I proactively held onto my papers before starting this video.