AI’s 48-Hour Frenzy: Breakthroughs, Billion-Dollar Bets & Bold Policies (July 22–23, 2025)
Google and OpenAI won gold medals at the International Mathematical Olympiad by solving 5 of 6 problems with new general-purpose reasoning based on natural language understanding. OpenAI disclosed its experimental model scaled up test-time compute to think longer and in parallel, a very expensive undertaking. Google DeepMind’s Gemini Deep Think, unveiled in May, solved IMO problems within the 4.5-hour limit entirely in natural language. MIT CSAIL and Facebook AI researchers presented ICML 2025 work showing image generation with encoder-only tokens, enabling inpainting without a generator and morphing a red panda into a tiger. Alibaba’s Qwen-3 Coder, open-sourced for software development,