Welcome to Jumble, your go-to source for AI news updates. This week OpenAI shipped its first piece of hardware, and it is not the pocket gadget everyone expected, while Moonshot labs dropped a model that just muscled past the frontier. Let’s dive in ⬇️
In today’s newsletter:
🖥️ OpenAI's first gadget hits desks
💥 Kimi K3 challenges frontier AI labs
🏛️ Xi's global AI rulebook
🎥 Google puts you in the video
🎯 Weekly Challenge: Kimi vs your usual AI
⌨️ OpenAI's First Physical Device Is a Keyboard
OpenAI entered the hardware market this week, and its debut is a $230 light-up keyboard for people running fleets of AI coding agents, not the pocket companion the rumor mill promised.
Would you pay $230 for a keyboard that runs your AI agents?
ㇽ️ A Command Center for Codex
The Codex Micro, made with keyboard shop Work Louder, has keys that light up to show agent status and a dial that sets how hard an agent thinks. OpenAI is pitching it as a limited-run desktop hub for agentic work.
🍎 The Bigger Device Is Still Coming
The flashy keypad is a warm-up for a screenless, portable speaker with moving parts, reportedly designed by former Apple engineers. Apple is now suing OpenAI, alleging its leaders lifted confidential hardware secrets.
🐲 Kimi K3 Crashes the Frontier Party
Moonshot released Kimi K3 on Thursday, a 2.8 trillion parameter open-weight model that testers rank right alongside the best from OpenAI and Anthropic, at a fraction of the price.
🥇 Beating the Big Labs at Their Game
In blind tests, developers preferred Kimi over every leading U.S. model for front-end coding, including Anthropic's Fable 5 and OpenAI's GPT-5.6 Sol. On a broader ranking it slipped past Anthropic's Opus 4.8.
💵 A Direct Shot at U.S. Pricing
Because Moonshot charges well below the premium labs, it is forcing a hard question about how long American companies can keep top dollar prices. Rival DeepSeek is expected to answer with its own upgrade any day now.
Weekly Scoop 🍦
🎯 Weekly Challenge: Put Kimi Against Your Usual AI
Challenge: Kimi K3 just outscored the biggest labs, so this week you get to test the hype yourself and see if it holds up on real work.
Here's what to do:
🔎 Step 1: Find Kimi The easiest way in is kimi.com in your browser, no download needed. If that is slow, grab the Kimi app from your phone's app store, or use a free browser playground like OpenRouter that lists K3.
🆚 Step 2: Give both the same job Pick one real task, like turning a messy block of notes into a clean follow-up email, and paste the exact same prompt into Kimi and into ChatGPT, Gemini, or Claude.
📊 Step 3: Judge the results Read both answers side by side. Which one nailed the tone, caught the details, and needed the least cleanup? Reply and tell us whether the free model actually kept pace.
Would you trust a $230 keyboard to run your workday? And now that a free Chinese model is beating the biggest labs, does brand loyalty in AI even matter anymore? See you next time! 🚀
Stay informed, stay curious, and stay ahead with Jumble!
Zoe from Jumble
