So happy that @numerai is finally getting the recognition it deserves. Met @richardcraib back in 2019 and was captivated by his visionary concept. Knew it then, but this has been a long time coming. Congrats to the team!

We’re officially opening up access to SAI – a new platform for reinforcement learning. Submit models, compete in structured challenges, and learn from others — anytime, not just during a conference. Try out SAI now -> 🌐 Documentation -> 📄

Meta is offering $100M+ packages to lure OpenAI’s top minds. And that's not even working. Because AGI isn’t just another product race. It’s the highest-stakes competition on the planet. The opportunity cost of not being the first is staggering. You should be paying attention.

"Arbitrum doesn't have any games" Shut up. Here are 10 arbitrum games you can jump into right now, with gameplay ⬇️

Don’t think this is impressive? Look again. Flipping a box might seem trivial, but this robot chose to do it. No scripts. No remote control. Just a system that saw the task and figured it out. It spotted the barcode. Knew it had to face down. Flattened the package to help the scanner. These aren’t memorized steps. This is real understanding. When it starts to feel normal, that’s how you know it’s working. These are the first steps.

We’ve written about this previously. LLM = Artificial general knowledge LLM =/= AGI. In order to achieve AGI, you need 👇

looks pretty insane tbh

A new playground for RL. Coming soon with SAI!

In 2013, this would've been a Black Mirror episode. In 2025, it's just another day at the mall. This is how it happens. One robot dog stroll at a time.

Video generation went from Temu Will Smith eating spaghetti to cinematic realism with Google's Veo 3 in two years. Robotics isn’t there yet. Over the next few months, you’ll see a flood of robotic demo videos. Most will be tightly-scripted, carefully-lit, and overly-choreographed to sell hype. This is the spaghetti phase. At NRN, we’re building towards the Veo3 moment for robotics. Real-world systems that adapt, compete, and scale. No tricks. No heuristics. Just robots that can actually hold their own.