7/ I phrased the ergodicity problem as the difference between if and when. In some cases we just have to wait, and it will happen. Elsewhere, whether it happens depends on immutable facts. Individuals can sample whens; only collectives can sample ifs.

three multibaggers meet a dystopian llm...

“the models we have remain, at their core, next-token roulette wheels. Chain enough spins together and tiny error probabilities compound into existential glitches.”

inverting Taleb's "when you see fraud..." famous quip, I get this, "if you see virtue and you say virtue, you are not virtuous." a "not yet" would better align the latter with Kris's up-and-to-the-right vector-like post: moontowermeta.com/the-devils-tri…

best cautionary tale I've read on bitcoin in a while h/t Quoth the Raven quoththeraven.substack.com/p/crypto-will-…

"all of the leading [AI] agents are essentially while loops with llm + tool calls" —Ankur Goyal

In thinking through future implications of AI, it is more important than ever to be able to hold multiple opposing ideas in your head at once. Because the probability distribution for any outcome has widened and flattened, dramatically in many cases.

Earlier in August, Amplitude announced Q2 2025 earnings results. We delivered the highest net-new ARR in 11 quarters years. This reflects years of focus on the enterprise. Here's a recap of the Q&A portion of our call.

cc Spenser Skates

Adam Smith's Backdoor, a Cassandra-like post by Kris moontower.substack.com/p/adam-smiths-…

Peter D. Kaufman's win-win compounding framework

Latest JPM CIO Survey - CIOs primarily looking to "buy" AI agents from SaaS/Apps providers vs "build" internally...(and have ample budget)

the question is not why, but when will France—a country conquered by its administration—fall? good post by Marc Le Chevallier engelsbergideas.com/notebook/stati…