Nag Murty

A difference in company philosophy: @neuralink: put wires on brain. @CorticalLabs: grow brain on wires. Cortical Labs just completely changed my dreams and nightmares. Here it is in Hon Weng’s hotel room. He is showing this off tomorrow at a brain conference in San Francisco. You get a sneak peak tonight.

someone please start a sirius cybernetics corp so we can get robots with GPPs.

We've built a simulated driving agent that we trained on 1.6 billion km of driving with no human data. It is SOTA on every planning benchmark we tried. In self-play, it goes 20 years between collisions.

Tomorrow we're finally unveiling our company, it's what I've wanted to build since I was 9. Until then, here's a thread of my favorite tweets from the premiere/movie today — for memories :)

"Robotics companies now are where LLM companies were five years ago. Sometimes in the next five years we will see the ChatGPT moment for robots." —@bobmcgrewai

those who think RL use less compute don’t know RL at all 😅 SFT: human generates data and machine learns RL: machine generates data and machine learns

Deepseek V3 and R1 discourse boils down to this. Shifting the curve means you build more and scale more dummies

タコにヒントを得た対数螺旋型マニピュレータ 多種多様な物体を扱うことができる youtu.be/2GFyFmMm9-A #manipulator #gripper #RoboticHand #biorobotics #biomimicry #バイオミミクリー #生物模倣 #octopus #SpiRobs

This is a brilliant chart of China's tech-industrial ecosystems, by Kyle Chan. Most of us economists don't understand that clusters of technological capabilities cut across our categories of 'sectors' and 'industries'. This chart gets that across in a fantastic way.

@Scobleizer its for RL at scale. the holy grail is robust reasoning, not knowledge.

The thing I don't get about Stargate is where is the data? Let's assume funding is secured (a big assumption) and that you can build a super huge datacenter with 100s of thousands of NVIDIA cards and an Oracle database server or two too. Where is the data coming from to keep the datacenter busy? I get how Tesla and X data keep Elon's datacenters busy. I get how a couple hundred million users of OpenAI's ChatGPT keep its data centers busy. I get how YouTube's data would keep a Google datacenter busy. But I don't see where the big dataset is coming from for the Stargate fund to use. So what really is at play here? 1. Hosting a super huge military simulator? 2. Getting all small businesses in USA to use AI? (That's the best outcome, but would require radical education of everyone running small businesses). 3. Hosting all of Walmart or all of Las Vegas? (Steve Wynn's datacenter is a fraction the size of eBay, which is located in the cage next door at Switch's datacenters in Las Vegas). 4. All Apple's data? Who has a big dataset that could keep a super huge datacenter busy? Something doesn't smell right here. @softbank usually doesn't fund charities. So, there must be a business reason to open up such a huge fund. Are they hoping to wipe out @ycombinator and fund every cool AI company that gets born from now on? That won't work either, because YCombinator brings a lot more to the table than Softbank does. But with AI you must always follow the data, not the money. The data is what matters and so far I don't see a real dataset involved here (IE one that has hundreds or thousands of petabytes that isn't already being served well by existing datacenters, or that would greatly grow in value once shoved through a bunch of NVIDIA cards). Do you?

@chris_j_paxton agree its awesome. Its just how capital expectations have been conditioned by the speed and ease with which information startups scale. Bit seems to be more lucrative than It.

@MurtyNag Waymo is pretty awesome though

the curse of robotics is that capital expects a Google, but physics allows a Waymo.

this is so cool!

IMHO, the problem with all the various "LLM Agents" writeups/tutorials floating about is that they all think agency can be defined in a hand-wavy fashion without the nitty gritty of actions, state, environment, models, planning, execution etc. Unfortunately it doesn't quite work.

over the coming decades - humanoid robots are going to age in reverse. they’re like frail geriatrics today - but over time should become nimbler and faster.

human brains are capable of forgetting just as quickly as they learn. Is this is a feature, and not a bug?

respectfully disagree - I think we're confusing capability with robustness. This makes for impressive demos but doesn't solve for product. Current VLM approaches can show great generalisation - but robotics is a problem of robustness/ high 9s reliability, done using low latency, high efficiency compute. And unlike software agents - you cannot "human in the loop" robotics to the same extent to backstop false positives.

If you want to see why we’re finally in for a robotics revolution, just play with the video mode in ChatGPT. It becomes instantly obvious how much visual capability is packed into these models.