I just used up my entire usage limit on #claude with a single prompt and got no result.
I was using the claude #agent via #ACP in the #zed IDE with claude-sonnet-4-6. The prompt stated to use 4 big pdfs to read and implement code based on it. It said it wrote code in the thinking parts but actually didn't. Then it hit the limit after maybe 30 minutes. The context window of claude #sonnet is 200.000 tokens but it only needed half of that to exhaust the usage limit.
This shit can't be real. They gotta rugpull the stock market quickly or go bust.
