Does the calendar say October? Well, like the good comptroller I am, I waited for the period to end before reviewing the books.
I don’t normally do that? You didn’t know I was a comptroller? Okay, fine, look, I could have posted this last week, but I had two good reasons not to.
First, last week’s Autofac Era essay was ready to go. It came together quickly after sitting on the ideas for a couple months, and I was excited to get it out.
Second, I was very close to finishing revisions on Chapter 4 of the book, and I’m happy to say they’re now complete! They took a bit longer than expected, both because the original version of the chapter had more factual mistakes than I had realized that needed fixing, and because I had to completely rewrite one of the sections based on the feedback that, to be correct, I’d really need to talk about the Löbian obstacle instead of issues with Bayesian self trust.
That meant I had to understand the Löbian obstacle well enough to write about it, which ate up plenty of time because I didn’t understand Löb’s theorem as well as I needed to. Much thanks to Claude for being smart enough to be able to read the original papers and keep pointing out my mistakes as I tried to explain the ideas by just continually checking if my words literally made sense compared to what was in the papers.
As ever, I’m hopeful that revisions on future chapters go more quickly. We shall see.
Other Happenings
I’ll be at the Curve conference this weekend in Berkeley. If you see me there, say hi!
I’m giving a talk about the origins of the post-rationalists later this month. Depending on how it goes, I might write something up here, too.
I have more posts about AI in the pipeline. Make sure to subscribe if you want to get those as soon as they are released.
And if you really want to know “What’s happening?”, you should follow me on X, the everything app, where I sometimes post interesting things that go under appreciated.
