Hook
More breakout videos from this creator.
The rumors are true. Anthropic and Open AI designed their models to get dumber the more you use them. You've definitely felt this. You're going back and forth with Claude or ChatGPT, and after a while, it starts getting forgetting things or just bugging out. Researchers found models get up to 39% dumber the longer a conversation goes. The official term for this is semantic degradation or context rot. But there's a simple fix. I'm Ben Gusberg. I'm a Stanford AI startup founder. We've raised $4 million. The fix is called context transfer. Open a new chat in Claude or ChatGPT and drop in this exact prompt. This will create a skill for you that summarizes all the key context from a conversation thread. Every goal, task, person, link, file. Then, anytime you notice GPT getting dumber, just type slash context transfer and you'll get a single text block that you can just copy and paste into a fresh chat and continue the conversation with nothing lost. This solves the context rot problem and allows you to start again with all the same context. Try this out and let me know what you build. I'm Ben Gusberg. Follow me for more founder life and leadership.