Unless you have access to the same compute limits as the guy at Anthropic, and the internal model they use, you probably would just have spent a lot of tokens without success. So you can take solace in that.