I made AI models play badminton to find out which one can reason fast AND accurately 🏸
And I was shocked: OpenAI's own open-source model GPT-OSS-120B scored higher than GPT-6 Luna and GPT-5.6 Sol. How it works 👇 Each AI gets the game state as numbers: • the ball's x/y
"I say what everyone wants
Fucking around freely is now called art!” (Tools: suno, deepseek, comfyui) LoRA trained specifically for Re-l Mayer.
In "The Big Fat Fish of the Xiaonan Liang Family"
In "The Big Fat Fish of the Xiaonan Liang Family", when all the dragon maids become AI girls - even if the server is busy, the OP still has to be completed. (Contains a large number of Silicon Valley easter eggs such as Lao Huang, Ultraman, and Musk, see how many you can recognize) Mr. Liang's Whale Maid (Miss Kobayashi's Dragon Maid OP - All AI Waifu Ver.)
People paying for the biggest model on the market and still getting cones on tables
Meanwhile deepseek v4.1 flash, 3 follow-ups: full paris, 1888, tower mid-construction, the seine, bridges, waterfalls it was never the price. it was the prompting
You have reached the end of the archive
All of deepseek