B-Rabbit

@crystalsssup @JingyuanLiu123 @grok what is Muon optimizer

@JingyuanLiu123 Deepseek V4 employed Kimi-style's Muon Optimizer😭 Muon optimizer is validated at scale by Kimi & Deepseek

Just imagine for a second that China with its open source was not in the game. How the world would look right now

With Kimi 2 k2.6 and deepseek v4 pro, I am more sure of this post

Only possible way for @cursor_ai to survive is that the open source community closes the gap with @OpenAI and @AnthropicAI on coding.

People are brute forcing any code. So the developers want the best model because they do not check their code anymore. In reality, if anyone creating any product, they would be okay with a model 50-80% cheaper but performing 20% worse. Low-cost open source will triumph.

@icanvardar Best AI company open sources. So I would with either Kimi or Deepseek in general and Google in the West.

best ai company?

@bridgemindai What u mean?

GPT 5.5 xHigh passes the Car Wash Test. Finally. A model that can think.

@sdlangshaw @Kimi_Moonshot What do you think that we are trying to convince you of here?

@4SimonSays @Kimi_Moonshot Not very convincing output.

Meet Kimi K2.6: Advancing Open-Source Coding 🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench Multilingual (76.7), BrowseComp (83.2), Toolathlon (50.0), Charxiv w/ python(86.7), Math Vision w/ python (93.2) What's new: 🔹Long-horizon coding - 4,000+ tool calls, over 12 hours of continuous execution, with generalization across languages (Rust, Go, Python) and tasks (frontend, devops, perf optimization). 🔹Motion-rich frontend - Videos in hero sections, WebGL shaders, GSAP + Framer Motion, Three.js 3D. 🔹Agent Swarms, elevated - 300 parallel sub-agents × 4,000 steps per run (up from K2.5's 100 / 1,500). One prompt, 100+ files. 🔹Proactive Agents - K2.6 model powers OpenClaw, Hermes Agent, etc for 24/7 autonomous ops. 🔹Claw Groups (research preview) - bring your own agents, command your friends', bots & humans in the loop. - K2.6 is now live on kimi.com in chat mode and agent mode. For production-grade coding, pair K2.6 with Kimi Code: kimi.com/code - 🔗 API: platform.moonshot.ai 🔗 Tech blog: kimi.com/blog/kimi-k2-6 🔗 Weights & code: huggingface.co/moonshotai/Kim…

@OpenAI I am sticking with Kimi this time

Introducing GPT-5.5 A new class of intelligence for real work and powering agents, built to understand complex goals, use tools, check its work, and carry more tasks through to completion. It marks a new way of getting computer work done. Now available in ChatGPT and Codex.

@ClaudeDevs We know 5.5 is coming but do not leave we listened you fixed it and here more usage, if you go to codex that extra usage will be wasted

Over the past month, some of you reported Claude Code's quality had slipped. We investigated, and published a post-mortem on the three issues we found. All are fixed in v2.1.116+ and we’ve reset usage limits for all subscribers.

@DeryaTR_ Would have make more sense if you did not decode it

NS41 as a base64-encoded string, yields the bytes for the text “5.5” Wonder what it means? 😊

@remarks @grok what are other drugs that are classified as Schedule III

Kimi claw group chat is sick

@rsuyoy 2 years

What's a reasonable timeline for a GPT 5.4/Opus 4.6 level LLM that runs locally on a good Macbook 10 years? much less? much more?

@Kimi_Moonshot are you guys gonna create cloud version of kimi code like codex, cursor or claude code?

@AdrianDittmann I guess the will be the fourth? So being “one of the best” is pretty easy i hope he becomes the best

John Ternus will be one of Apple’s greatest CEOs

@BenjaminDEKR @xai You would like that

Noteworthy: the SpaceX - Cursor announcement has "xAI" NOWHERE in the post. So far @xai hasn't posted about it, nor is it on the xAI blog. This seems like a very purposeful choice. xAI brand being slowly phased out?

@AshCrypto @grok double check

THIS IS MASSIVE FOR CRYPTO 🇷🇺 Russia just "PASSED" the crypto regulation bill to allow businesses and companies to use crypto as payment for cross-border and foreign trade settlements, even under sanctions. $BTC and $ETH are expected to be the first assets approved under the framework

@BenjaminDEKR No it doesn’t

Cursor is specifically saying the partnership is to improve COMPOSER (their in-house AI code model) NOT to improve Grok or other xAI / SpaceX models This seems so backwards and strange

@NoLimitGains @figma and thats on them

🚨 Meet the first company officially wiped out by AI. Which one is next?

@BenjaminDEKR dude leave them alone, GOD…

Is Cursor really worth 4% of SpaceX's entire (projected) IPO value?