Diego Prats | 🤖

@OpenAI @RoboPapers @nvidia @GoogleDeepMind 3. 1X World Models by @1X_Technologies - train a world model on real humanoid data to imagine futures and evaluate policies without running the physical robot 1x.tech/discover/1x-wo… robopapers.substack.com/p/ep61-1x-worl…

@OpenAI @RoboPapers @nvidia 2. Gen2Act by @GoogleDeepMind - generate a video of a human doing the task, then condition the robot's policy on that generated rollout homangab.github.io/gen2act/ robopapers.substack.com/p/ep57-learnin…

I am not sure it is appreciated how much of the @OpenAI Sora shuttering is actually a reallocation of video-generating resources away from entertainment and toward physical AI. Even though they literally said: "the Sora research team continues to focus on world simulation research to advance robotics." 🧵

Today, we’re releasing Migas 1.5: the first foundation model to fuse text and time series. Until now, forecasting models have only looked at historical numbers. Migas 1.5 changes that by letting users incorporate real-world context directly into the forecast. This enables teams to forecast with essential context like earnings reports, policy changes, market events, supply shocks, and more. This directly results in more accurate forecasting and enables complex scenario analysis, especially when historical data is sparse. Highlights: - Highest Elo rating against leading foundation models on 86 real-world datasets - 75%+ win rate against all baselines (even Migas 1.0!) - Up to 14.2% lower MAE in short-context forecasting - Fully open source - Premade Claude skill to get you started in seconds We’re excited to open-source Migas 1.5 and eager to see what the community builds with it. Links in comments.

A favorite parlor game among fellow parents in tech is “what should I tell my kids to do/study?” I do not know, but I can tell you an observation (not recommendation): 1. I have NEVER met anyone in Silicon Valley whose life would have been better by knowing 10-50% more math, physics, or CS. I never met the person who because they knew a bit more calculus, they are truly happier or more successful. 2. I have met MANY people where 5-10% more discipline, patience, curiosity, grit was the difference between a great life, a good one, or painful one. And I have seen AI only sharpen the divide.

Live demo from Generalist AI at gtc -- inserting phone into case. Next to it you can try out doing the task yourself with a Universal Robots leader/follower setup. Its harder than the robot makes it look! Seeing demos like this live and in person always hits different

@sohan_zhang @HapticLabsAI 🤣

@mexitlan @HapticLabsAI following hapticlabs to help pump the growth rate 🔥

We did it you guys. Since posting yesterday, @HapticLabsAI went from 3 followers to 24. 800% growth. Assuming a CONSERVATIVE 50% slowdown every day, we should reach the entire human population of Earth in about 16 days. Agent population may outpace it. youtube.com/watch?v=e6LOWK…

Thank you to @TheCbac @chris_j_paxton @ai (what a crazy handle right) for the help!

My pet peeve with startup podcasts is they are usually: 1. How startup X is amazing 2. Why tool Y is amazing I find these about 50% inspiring, 25% envy, 25% depressing. What I’ve always wanted instead are real problems to chew on. So as an opening salvo for my startup @HapticLabsAI, I wanted to start with exactly that for my fellow builders.

At BASS and SBC conversations around AI and crypto circled the same question: How do we trust systems we can’t see inside? @alexanderfowler and @danielmarinq argued that trust will come from infrastructure designed for proofs. Read more: blog.nexus.xyz/verifiable-ai-…

Tanner is awesome

🚀

🚀🚀