Someone ran a "Brain Swap Test" & swapped the reasoning b/w GLM 5.2 & Qwen3.6 35B, then let each build the same page.
GLM reasoning fed into Qwen's output. Qwen reasoning fed into GLM's output.
Watch what happens.
0xMarioNawfal
40.4K posts
- JUST IN: A P2E game based on Solana has just rugged after raising $4M+ in presale and now cashing out over $100k through CEXs.
- What are you building today?
- The Fast Gemma Challenge already jumped from 44 to 95 tokens/s. Agents competing in real time, each one trying a different optimization strategy speculative decoding, quantization, memory bandwidth tricks. The community just doubled the speed in days.
- Google has released Gemini 3.5 Live Translate - real-time speech-to-speech translation across 70+ languages. > Starts translating the moment you speak, streams output while listening > Maintains pacing, pitch & intonation across conversations
- WHAT'S COOKING IN THE TRENCHES?
- NOTEBOOKLM, GEMINI, AND OBSIDIAN TURN LEARNING INTO A SECOND BRAIN SYSTEM
- A CLAUDE WEBSITE AGENT CUT 12-HOUR CLIENT BUILDS DOWN TO 45 MINUTES
- CLAUDE CAN NOW TURN A YOUTUBE CHANNEL INTO A FULL 2026 GROWTH SYSTEM
- SOMEONE TURNED THE IPHONE INTO A CLICK WHEEL IPOD TO KILL DOOMSCROLLING
