Umm which local LLM (presumably it’s running on a laptop) can do even a fake-passable job at this? He is clearly not actually using a local model if this is how he’s generating his peer reviews, especially if he’s been doing it for some months.

Haunted Casio viable though somehow

Would you consider automating the posting of the thought + link from your blog onto Bluesky / Twitter?

Until more folks do what you do it’s hard to standup an RSS reader etc to follow along while also reading interesting links from Twitter.

Here's one counterargument: the way he describes economists reading articles is also how mathematicians read math papers. We look at the results to see if they're interesting to us, and only then at the proof method and computations if we think these generalize for our interest case.