Chief Business Officer at Humanoid | Driving the capital, business, and government relations behind reliable, safe, and helpful AI-powered humanoid robots
The world is messy, and robots don’t always operate in perfect conditions. Like humans, they sometimes make mistakes.
One of the things we teach our robots at Humanoid is how to correct and recover when something goes wrong. Here's how we do it 🧵
The AI Singularity
The argument goes like this:
1. Humans build an AGI.
2. The AGI becomes good at AI research.
3. It designs a smarter AI.
4. That smarter AI designs an even smarter AI.
5. The cycle repeats faster and faster.
Looking at the results and capabilities from the various labs over the past few weeks I would say we are firmly in this loop now.
The next 18months will be wild.
Recursive self improvement will dramatically increase capability very quickly from here.
Marginal costs of all models will go to ~$0.
I’ve never posted on X.
Last year I was at Department of Energy with an unusual view into what’s happening across AI, energy, and infrastructure. The old systems can’t handle what’s coming.
There’s just an insane amount we need to build in the physical world, and the opportunity and urgency are greater than ever.
Every able person should be contributing to the effort, and I’ve gotten back into the arena to do my part. More to come soon.
Why Sudoku?
It tests constraint satisfaction and LLMs can't do it.
LLMs are good at language problems. Robotics and industrial control systems problems aren't like that.
That's why we created a new architecture that is the foundation of Kona’s Energy-Based Reasoning Model.
@tbpn@ylecun Yann will be guiding our work as we develop systems that actually reason, and energy-based models moving from a research concept to real products is a fundamental step in that journey. Follow us and our Founder and CEO @evelovesolive for more!
https://t.co/fZffoFxSPM
Palmer Luckey says UAPs likely come from the distant past, understanding them will render current technology ‘irrelevant’
“There’s a lot we don’t understand about these vehicles — What they are, where or when they come from.”
“If and when we figure out what’s going on with UAPs, UFOs, that whole universe — Everything’s gonna change.”
Our Aleph prover agent just hit #1 on PutnamBench, a benchmark built from Putnam problems - one the hardest college-level math olympiad - fully formalized with machine-checked proofs and no human involvement. Putnam problems are often considered harder than IMO problems and span a wide range of topics, including calculus, number theory, group theory, and other core areas of mathematics.
This is strong evidence that AI can handle deep, multi-step reasoning with correctness guarantees — the same kind of technology we’re using to verify real software, hardware, and scientific discoveries that require formal logic.
https://t.co/vc15k0qbG2
We just tested Aleph prover on this version of Erdos #124 problem and were able to prove it in less that 2.5 hours and under $200 in cost: https://t.co/Q8boT8zOh2
https://t.co/xsyWv2ehDC
We are outsourcing perimeter security at Kabul airport to the Taliban. The Taliban has put the al Qaeda linked Haqqani network in charge of security in Kabul. So we’re depending on al Qaeda linked terrorists to provide security as we withdraw.