adam wik
186 posts
The first experimental evidence of recursive self-improvement (RSI). Autoresearching the autoresearch agent for eight days. The result beats the harness we hand-tuned for two years, on held-out benchmarks: 🧵(1/7)
We gave a coding agent a goal and a time budget: build a training environment and teach a vision model to count colored stars. Using autoresearch with NeMo RL, NeMo Gym, and reusable skills, the agent set up, trained and evaluated the model while the researcher steered the work. Qwen3-VL-2B went from 25% to 96.9% accuracy, and the agent even proposed the next experiment on its own.
I’ve never seen a white boy outwork a Mexican. Anyone who says otherwise has never been on a blue collar job-site in their entire life.
It’s bait
World Cup RO32 Prediction✅ 🇿🇦 South Africa 0-1 Canada 🇨🇦 🇧🇷 Brazil 2-1 Japan 🇯🇵 🇩🇪 Germany 3-1 Paraguay 🇵🇾 🇳🇱 Netherlands 3-1 Morocco 🇲🇦 🇨🇮 Ivory Coast 0-3 Norway 🇳🇴 🇫🇷 France 2-0 Sweden 🇸🇪 🇲🇽 Mexico 1-0 Ecuador 🇪🇨 🏴 England 2-0 DR Congo 🇨🇩 🇧🇪 Belgium 1-0 Senegal 🇸🇳 🇺🇸 United States 2-1 Bosnia & Herzegovina 🇧🇦 🇪🇸 Spain 3-1 Austria 🇦🇹 🇵🇹 Portugal 2-0 Croatia 🇭🇷 🇨🇭 Switzerland 1-2 Algeria 🇩🇿 🇦🇺 Australia 0-2 Egypt 🇪🇬 🇦🇷 Argentina 4-0 Cape Verde 🇨🇻 🇨🇴 Colombia 1-2 Ghana 🇬🇭
