Chatting with a good friend who has worked on various AI models while pursuing a PhD at CUHK, I gained the following insights: 1. OpenClaw is garbage. 2. A senior student from Beiyou used Claude to write a project for 2 weeks and sold it for 30 million dollars. He also wants to start a business but lacks ideas, and the window of opportunity is only until this year. 3. Claude already has the ability to train itself, and other large models basically have self-training capabilities as well, but no one is saying it publicly. 4. Video AI will see a qualitative leap in the next two years; editing AI is much harder than expected. 5. I am particularly anxious about losing my job.

Please don't scold me 🥲, I'm just sharing his thoughts and chat summary, and also 2 is 30 million RMB (he said it was in USD during the chat, I checked and wanted to make a correction.