building autonomous american robots @mecha_corp. founded @apolloaiapp (acquired by @liquidai).
San Francisco, CA
Joined March 2009
- Tweets9.9K
- Following1.2K
- Followers32.8K
- Likes119K
Mecha has a website now. Check it out at mechacorp.ai
Reach out if you’re interested in building the robots that will be deployed to cities, businesses, and homes.
Amazing.
Been running Opus 5.5 agents over the notebooks of Gauss, Riemann, Hilbert and Grothendieck. around 12,000 pages so far. Heres the best finding so far:
Riemann's 1860 shock wave paper is famous for getting the energy part wrong however his unpublished sheets at Göttingen have it right the Rankine Hugoniot conditions a decade or two before Rankine and Hugoniot plus results usually dated 19061937, all in his own notation. A second sheet in another folder has the same physics.
There are probably thousands of potential undiscovered stuff in the research papers of the smartest ever to live imo all of them should be made public and reviewed by ai years of work can probably be transcribed by agents in days
The iPhone Duo is amazing for managing agents and reviewing work. I cannot imagine going back to a regular phone.
7
53
4,174
This attitude is probably the most defining characteristic of the US, and one of the things I love the most. “Oh they’re banned terrestrially? Then it’s decided: we go to space.”
One of the glaring flaws in our system: nobody knows what they’re voting for.
Prop 38 reads to everyone as “should we cure cancer”. Therefore if you vote yes, you are good. If you vote no, you are evil.
The reality is everyone just voted to give 3 guys 4 billion dollars.
This is Snoopy, our sixth robot iteration in just the last few months. Built to handle terrain: Snoopy can climb, chase, recover, and monitor.
Autonomous robots will be the largest industry in the world. We have no plans to slow down. Come and visit us in the dogpatch to learn more.
It makes sense that this works, they're just weights -- but at the same time, holy shit? This is an intelligence improving another intelligence. The fact that Astra can do it so easily is astounding. It's like if we could look at mouse neurons and swap them around to make it talk.
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-sp…
github.com/RikiyaT/lm_und…
huggingface.co/Qwen/Qwen3.5-0…
huggingface.co/RikiyaT/Qwen3.…
Kurosawa’s Ran is a great example of the “film vs digital” look debate at its extreme.
No matter how much film emulation gets stacked on top, few modern digitally shot movies look as good as this one from 1985.
2
40
3,775
Apple apparently acquired a subvocalization company a month ago? Please add this into AirPods immediately @johnternus
2
9
1,281
I'm convinced AI & continuous dictation will lead to many "codex leicester-esque" tomes of work from unexpected places. There are many hidden geniuses out there that just need a secretary.
I wonder if this replaces the pencil for memory for us? Pencils helped our ability to think long and the calories per word were high. The act of writing was also an act of memory.
When you speak, you also activate different types of creativity. So, this should be interesting.
1
5
1,133
