Let's See If OpenAI's Deep Research Can Tell Us about the Science of Beeminder

OpenAI’s Deep Research tool was made available to premium users the other day and I tried it out with the following prompt:

“What does the science say about Beeminder’s efficacy?”

It asked a couple followup questions (Beeminder specifically? No. Any specific outcome measures in mind? Also no.) and then it thought for 6 minutes and churned out a report. I’m quoting it in its entirety (for Science!) with my reactions interspersed. But I’m mercifully collapsing most of it so you can mostly just read my comments and reactions. (The only edits I’ve made to the AI’s words is replacing repeat citations with “[ibid.]”.)

See the Beeminder blog for the rest.