On November 23, 2025, I criticized entrepreneurs and marketers who treat AI as a fad and see virtue in doing difficult tasks manually when AI could do them in seconds. My principle was “Efficiency > hard work,” and I imagined this moment might resemble the dot-com era.
On November 23, 2025, I said I had sent many back-and-forth messages from one LLM to another to get different opinions, and I liked the idea of putting that process in one place. I was reacting to @karpathy’s description of llm-council, not claiming it was my app: the described workflow sends a query through OpenRouter to multiple models, lets them review and rank one another’s anonymized responses, and then passes the material to a “Chairman LLM.” The captured description cuts off before explaining that final step. The listed model examples were openai/gpt-5.1, google/gemini-3-pro-preview, anthropic/claude-sonnet-4.5, and x-ai/grok-4.
On November 24, 2025, I requested, “Launch a task in OpenAI agent mode and read a book.” I did not identify the book, and the captured request does not establish that the task was launched or completed.
Related: multi model ai tools, the ai wars heat up, ai tools empower creatives