AI, Robotics, Control, RL 🤖

🌍
Joined February 2015
Long-term games > Short-term gains
473
/anatomy explain how inverse and forward kinematics work, based on my blog. check 👇 airo7.me/blog/robot-arm-kine…
Introducing Anatomy Describe a machine in one line. Claude brings it to life. - Interactive isometric wireframes you can rotate, inspect and explore - WebGL shaders for fire, water, plasma and more - Real 3D motion, with individual parts that rotate and move Install it as an agent skill. Fully open source.
4
134
. @elonmusk played 4D chess. OpenAI launches Dots. You go to dot.com It sends you to Grok Bot. 😂 @spacexai already owned the URL.
2
44
Claude Opus 5.5 made this video entirely with code. 3,000 years of Ethiopian history. 162 seconds. No cameras, no stock footage, not a single photo. Every frame, every note, written line by line. 🇪🇹 @minilek ኢትዮጵያ · Land of Origins 🎙 Voice: ElevenLabs
1
31
not every low point means failure. keep going!
4
want → struggle → obtain → temporary satisfaction → boredom → new want → struggle
7
Grok 4.7 is a beast on hardcore electrical and mechanical engineering. it actually reasons through the physics instead of waving at a textbook formula.
17
all you need is Schmidhuber. @SchmidhuberAI with Jev bringing calibrated confidence and RLCD into the spotlight, it's worth remembering: modeling prediction reliability with RL goes back to 1991. people.idsia.ch/~juergen/FKI…
22
reality does not begin where perception begins. there is a world before we name it, measure it, or understand it. perhaps the unknown is not empty. perhaps it is simply unmodeled.
7
humans need simplification to act, but reality doesn’t simplify itself for our convenience. reality is continuous; cognition discretizes it. reality is complex; cognition compresses it. reality is particular; cognition generalizes it. reality exceeds the model; the model remains useful. but what is reality, when every way we know it is already a model?
10
we don't live by reality directly. we live through models of reality. those models are the handles that make action possible, but the world they grasp is always larger than the model itself.
10
rationals are the practical handle. irrationals are the exact law.
10
turned that into a published skill. install the skill on your Hermes @NousResearch agent and get production access first try. sandbox → production access, DKIM/SPF, SMTP credential derivation, the whole playbook that got me out of the rejections, now open source. github.com/airo7/aws-ses-ski… clawhub.ai/airo7/skills/aws-…
shoutout to Hermes agent. @NousResearch tried multiple times to get AWS SES production access. kept getting rejected. gave my hermes agent limited IAM access, it handled DNS/DKIM checks, use-case wording, and the support case details autonomously. production access granted in 3 days. wild what a properly tooled agent can unblock.
1
31
learn parameters, optimise decisions.
7
shoutout to Hermes agent. @NousResearch tried multiple times to get AWS SES production access. kept getting rejected. gave my hermes agent limited IAM access, it handled DNS/DKIM checks, use-case wording, and the support case details autonomously. production access granted in 3 days. wild what a properly tooled agent can unblock.
1
1
101
kimi 🔥
Releasing the model weights and technical report of Kimi K3. Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window. New model architecture: 2.5x the intelligence per unit of compute, not just more params. Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale. Model weights: huggingface.co/moonshotai/Ki… Tech report: github.com/MoonshotAI/Kimi-K… Tech blog: kimi.com/blog/kimi-k3
26
do your company accounting and tax with @grok 4.5 it’s actually insane. amazing model. big up the @spacexai team @elonmusk 🔥
1
31