Columnist argues revenue metrics alone don't reveal winner between Anthropic and OpenAI in early AI market

A tech columnist who claimed Anthropic was winning against OpenAI based on revenue was challenged by a podcast host as offering "a very SV/Twitter bubble take, not yet backed by real data," prompting reflection on whether traditional financial metrics reveal the true competitive picture in AI.

The author suggests more revealing metrics would include daily token generation per company, the percentage of tokens monetized and average price per token, and compute utilization rates broken down by purpose (training, inference, internal experiments, research) — most of which remain undisclosed.

Anthropic appears to be "revenue-maxxing" by limiting tokens in its Claude subscription ($20-per-month), forcing users to pay extra to complete tasks, while OpenAI's Codex subscription allows longer use without hitting limits as steeply or expensively, making it a better user experience for non-professional developers.