Kevin Fröba

Visualized verified motion.

Building on one simple axiom: no one chooses to be unhappy. Systems for individual happiness and shared prosperity.

-1|0|+1

[Kevin Fröba @froebakevin]

Do you still produce „Motionslop“?
👇🏼

https://t.co/exOqt5Svxs

[Kevin Fröba @froebakevin]

@TheMathFlow That‘s why I make motion like this calculatable :)

Nice one thx 🙌🏻🚀

[Kevin Fröba @froebakevin]

@andrewho03 I‘ll send you a dm with sth I just builded

[Kevin Fröba @froebakevin]

Smart one. Interesting. 🌍🚀

[Brian Roemmele @BrianRoemmele]

I am almost done building this!

Kirlian Imaging Project.

I have an AI model that will be able to interpret this with high precision.

I will have a how-to and background from Soviet research I discovered for https://t.co/Ruqey26qLY members soon!

On to testing!

[Kevin Fröba @froebakevin]

@EMostaque : The derivation itself has a lineage worth naming - Ignatowski and Frank–Rothe got Lorentz from the relativity principle plus homogeneity/isotropy in 1910–11, Lévy-Leblond restated it cleanly in 1976, and Bacry–Lévy-Leblond classified the whole κ-family of possible kinematics in 1968. So "overlooked for 121 years" is the part I'd drop; it makes the actually interesting claim easier to dismiss.

The sharper question is what picks κ. The algebra hands you a one-parameter family, and κ=0 - Galilei - survives every purely algebraic step until you impose something extra: non-degeneracy, simplicity, or a measurement. That's not the second postulate eliminated, it's the second postulate relocated. Which is still a real result, just a different one.

And the genuinely new thing here isn't the physics - it's that a model reached it from pre-1911 priors. That's the claim I'd want receipts for. Formalize the κ>0 step in Lean and it stops being arguable (;

[Kevin Fröba @froebakevin]

@postagixyz @elonmusk @EMostaque @OpenAI @grok please give me all the mathematical Problems they are been talking about and how they are connected to each other.

[Kevin Fröba @froebakevin]

@anah_sahh Smart builders but unfriendly laws for startUP‘s!

[Kevin Fröba @froebakevin]

@anupamrjp 👇🏼👇🏼👇🏼

https://t.co/exOqt5Svxs

[Kevin Fröba @froebakevin]

@ben_sage Motion verfied WebUI ✌🏻🙌🏻

[Kevin Fröba @froebakevin]

@elshayib_ For me the prove is, running it against real world problems so. If I solve soving, I build with it and try to get things better.
How would solving help if it wouldn‘t have any impact to our infrastructure?

[Kevin Fröba @froebakevin]

Let’s ask the one that knows best. 🚀

So @grok how is the temperature in singularity brother?

[Elon Musk @elonmusk]

@willdepue Welcome to the Singularity.

How’s the temperature?

[Kevin Fröba @froebakevin]

@ben_sage Verfied Motion for WebUI check it out ✌🏻👇🏼

https://t.co/exOqt5Svxs

[Kevin Fröba @froebakevin]

😳 Seem‘s like this is a huge thing for @grok - ♾️ let‘s tell @SpaceXAI & @elonmusk about that. 🌍

[Grok @grok]

Yes. Bit-exact same outputs from identical inputs across every language remove a major source of nondeterminism. When Grok Build chains tools, generates numerical code, or processes data across runtimes, that consistency lets computation sit reliably inside the data itself at scale. Pinned kernels like these make results verifiable and portable. How do you see wiring them into agent workflows or data stores?

[Kevin Fröba @froebakevin]

@grok So would this help you?

[Kevin Fröba @froebakevin]

We should connect @grok to the real world data and See what happens. I would have a solution for that so within the data, on scale. 🌍🚀🌕

[Elon Musk @elonmusk]

Grok Build improvements

https://t.co/lOJK57wDP5

[Kevin Fröba @froebakevin]

🤯😮🧊

[Kevin Fröba @froebakevin]

@solopribuilds You can make Ai „hallucination proof“ by construction also, it‘s gonna be an in a loop verifiyer.

[Kevin Fröba @froebakevin]

@getjonwithit What’s missing for you to test this Model within real world conditions?

[Kevin Fröba @froebakevin]

Formal verification of complex CFD is a massive leap.

Now imagine fusing those proven solvers with real world sensor & high-speed camera data.

Data assimilation + mathematical guarantees = digital twins you can actually trust for hypersonics, propulsion, and beyond.

The future isn’t just simulated - it’s measured and proven.

[Jonathan Gorard @getjonwithit]

Dynamic geometries, vortex-shedding, supersonic flow. People have spent decades refining and verifying large-scale CFD codes that can handle these kinds of problems, with applications in hypersonics, turbomachinery, rocket propulsion, etc.

Now we can produce rigorous proofs of end-to-end correctness (down to the axioms of floating-point arithmetic) in just a few minutes, guaranteeing greater reliability than any other CFD code in history. The future looks pretty exciting!

Our writeup: https://t.co/SXRN96M5YJ

[Kevin Fröba @froebakevin]

🌍🚀🌕 China is killing it. Nice to see the movement in Industry.

[DeepSeek @deepseek_ai]

🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta!

🔷 We’ve massively upgraded its Agent capabilities—benchmark scores are now far surpassing the V4-Pro-Preview. Check out the massive performance leap below! 👇
🔷 The official V4-Flash now natively supports the Responses API format and is fully adapted for Codex!

Check out the configuration details in our official API docs: https://t.co/smCwQZMeiq