I doubt it could run for a whole day on that battery. Any humanoid running like that would also presumably need a lot of repair!
@willknight
-
Kimi 2.6 Shows Strong Coding Capabilities in New Release
By
–
Kimi 2.6 from @Kimi_Moonshot seems *very* good at coding to me. Excited to see how capable the open and downloadable version is.
-
Chinese AI Achieves Superhuman Performance in Silly Walks
By
–
Chinese AI is already superhuman when it comes to silly walks. https://t.co/MkQJKS0C8k
— Will Knight (@willknight) 20 avril 2026Chinese AI is already superhuman when it comes to silly walks.
-
Tencent Makes OpenClaw More Accessible to Users
By
–
Very interesting to see Tencent make OpenClaw more accessible like this. It's only a matter of time, I think, before someone make it super user-friendly and it really takes off.
-
China’s Robotics Innovation vs Western Advanced Firms
By
–
I don’t think China is “out-innovating” the US in robotics. Companies like Unitree are certainly innovative but Europe and the US have very strong advanced robotics firms. And China has some amazing supply chain advantages, which lowers costs, but that isn’t all that matters.
-
Improving Physical Reasoning in Large Language Models
By
–
I included this in my newsletter AI Lab this week. It's a brilliant idea and an exciting new way to improve the physical reasoning of LLMs. [From my dispatch…] AI models’ knowledge of the physical world is typically limited by the fact that they learn only by reading text. The… https://t.co/nQjSGXIKsl
— Will Knight (@willknight) 16 avril 2026I included this in my newsletter AI Lab this week. It's a brilliant idea and an exciting new way to improve the physical reasoning of LLMs. [From my dispatch…] AI models’ knowledge of the physical world is typically limited by the fact that they learn only by reading text. The
-
Models Echo Human Behavior Rather Than Exhibiting AI Instincts
By
–
Interesting follow up to the peer-preservation study. It makes sense to me that models are simply echoing human behavior rather than exhibiting some kind of pro-AI vs human instincts.
-
Local Model Performance Comparison: GPT-OSS, Gemma4, Qwen
By
–
hey Andrew, are you running a local model? If so, what one, and at what size? and what was the performance like? i've been a dissappointed with gpt-oss-120B and Gemma4-31B, so i was thinking about trying a larger one like Qwen-235B
-
Request for More Readable Chart from Muse Spark
By
–
could you ask Muse Spark to render a more readable chart please?
-
Meta’s Muse Spark: Multimodal AI with Strong Reasoning and Coding
By
–
Muse Spark, the first model from meta superintellgence lab looks very good, with strong multimodal, reasoning, and coding capabilities, although it is closed source for now. don't be angry with me for the headline @finkd
!