Frank Hecker

Good general advice on using R productively.

The log-normal distribution strikes again, the time for times to task completion.

An analysis of project earnings on Patreon, a service with high-quality data available. Result: It’s the log-normal distribution.

One of the better “contra AI” articles I’ve read. Interesting that it closes with a poem, because poetry is one of the better examples of intelligent compression around, along with metaphors. Possibly misquoting Randall Jarrell on Wallace Stevens more abstract poems: “It is G.E. Moore at the spinet”. An LLM can explain it to you, but it hits with more force if you’ve read Jarrell, Stevens, and Moore, and have ever listened to a spinet. (I can check 1, 2, and 4, but not 3.) Compression is most powerful when you have some idea of what’s being compressed.

More on the theme of widening inequality in a post-AI world.