You are right to pushback, I did not provide a clear explanation 😂
Ve videu jsem to neviděl, ale hádám, že díry je možné vyplnit. Našel jsem zmínky třeba tady: https://t.co/q46NiMqsEW
Potom pomažeš "černou pastou" a zase vypečeš?
Nebo to možná někde vycvakli zazadu.
I gave xAI a lot of shit for Grok Build’s repo-upload behavior, so they deserve the same volume when they fix it.
I pulled the new signed release, planted fake secrets, and captured its HTTPS traffic with MITM proxies. The fix held as retention was off on first login and had no /v1/storage, GCS traffic or Git bundle, and none of my canaries left the machine.
Telemetry still runs, but I found no prompts, code, filenames, or secrets inside it. They published the source and the privacy-gate tests passed 5/5. Attaching my investigation here https://t.co/xc6hv6E81g
Max respect to the team for taking accountability and making a verifiable fix.
This benchmark uses ELO-style ranking with blind human voting. Plotting across a wider range shows that they are all good models :)
Same way top 20 GMs are all great chess players.
If you run Kimi K3 and Fable 5 head to head at their ELOs, you get 57/43 win rate.
Big news: Kimi-K3 by @Kimi_Moonshot is now #1 in the Frontend Code Arena with 1679 pts, surpassing Claude Fable 5.
This is a 17-place jump from Kimi-k2.6 (#18 -> #1).
In Frontend, Kimi-K3 ranked #1 in 6 of 7 domains: Brand & Marketing, Reference-Based Design, Data & Analytics, Consumer Product, Simulations, and Content Creation Tools, landing #2 only in Gaming behind Fable 5.
The full model weights will be released by July 27.
Congrats to the @Kimi_Moonshot team on this major milestone!
@fineri Hey Iain! This project was a big inspiration when we were trading ignition systems on the Firebolt @daretudelft. In the end we went with a pyro charge, but it was a close trade. I spent two weeks reverse engineering your igniter! 😁 Hope you’re doing well :)
Is Grok Build actually good?
It has the best TUI by far and the rate of improvement is impressive. But the model sucks compared to Opus.
It's like a lawn mower engine in a Ferrari.
Excited to see Grok 4.5 soon!
Grok 4.5, based on our 1.5T V9 foundation model, with Cursor data added in supplemental training, is now in private beta at SpaceX & Tesla. Early evals show performance close to, perhaps exceeding Opus.
RL is continuing to significantly improve the model, and the Grok Build harness gets better every day.
Nice work by all those involved!
Completely trained from scratch new models will be released by @SpaceX every month this year.
15 let součástí evropské vesmírné rodiny🚀
Před 15 lety vstoupila 🇨🇿 do Evropské kosmické agentury a začala nová éra kosmických aktivit ČR, zaměřená na rozvoj kosmického průmyslu🔧, pozorování 🌍 , průzkum vesmíru 🛰️🔭 atd.
Děkujeme všem, co se na projektech v @esa podílíte! 🤝