I am going to be honest with you. The first time I used Claude, I thought it was worse than ChatGPT. Slower, more cautious, less impressive. ChatGPT gave answers that sounded brilliant in 3 seconds flat. Claude took its time and sometimes said I am not sure about this.
That was July 2023, the day Anthropic opened Claude 2 to the public. Todd and I signed up that same afternoon. We had already used ChatGPT for 8 months and built over 3,000 custom agents, so we had a solid point of comparison. Almost three years later, Claude is the backbone of what we deliver: marketing strategies, bilingual content, follow up sequences, conversion funnels and complete digital ecosystems across more than 12 industries.
01 · Our timeline with Claude
02 · ChatGPT versus Claude after 3,000 agents
We use both, every day. We are not the X is better than Y crowd.
03 · The 63% discovery and the architecture mistake
Three months in, conversations were degrading and Claude gave three paragraphs of preamble before a simple answer. It was not Claude, it was our setup. A developer published an open source project, claude-token-efficient, with 8 rules that cut output tokens by 63 percent with zero information loss.
04 · The numbers (Todd's technical take)
05 · 6 things you can fix right now
Sources
Drona Gangarapu, claude-token-efficient open source project (github.com/drona23/claude-token-efficient).
