#hypothesis
The recursive hypothesis continues to be the threshold where people leap from the science to the religion of AI.
As Anthropic heads for a $1.4 trillion IPO on the promise of self-improving AI, Eryk Salvaggio asks what, precisely, is the “self” in recursive self-improvement — and whether the evidence supports the idea that we are on the verge of an "intelligence explosion."
Confronting the Limits of Recursive Self-Improvement
Eryk Salvaggio asks, what, precisely, is the “self” in "recursive self-improvement?"
www.techpolicy.press
November 17, 2026 at 4:18 PM
I'm not sure. We should test this hypothesis further.
November 16, 2026 at 4:48 AM
theyre calling it the tranny cavewoman hypothesis
across all eras of recorded history AND within the fossil record there is CLEARLY some subset of humans who modify their bodies with the tools and technology available to them and there is zero reason to be certain there weren't social gender reasons behind some of it even 100,000, 200,000 years ago
November 17, 2026 at 6:41 PM
The project began with a clear hypothesis and has developed into a rich descriptive account
November 16, 2026 at 12:44 PM
Let's test the hypothesis that Elon still thinks so.
November 15, 2026 at 6:28 AM
Fascinating hypothesis, have you tested it recently?
November 16, 2026 at 9:35 PM
He’s a clueless dumbass but based on what we know it’s a perfectly reasonable hypothesis.
November 17, 2026 at 1:51 PM
Her Wikipedia page doesn't seem to support that hypothesis.

en.wikipedia.org/wiki/Liz_Fon...
Liz Fong-Jones - Wikipedia
en.wikipedia.org
November 18, 2026 at 6:36 PM
Very plausible hypothesis on the NEET problem: underfunded students are doing the work instead. So they don't have time to study while others can't get work. A lose-lose situation.
Alan Milburn’s review of young people and work reports tomorrow. Ahead of it, a tale of six charts on what NEET counts can’t see – students 🧵

Among 18–21s, full-time students’ employment rate rose 2022–25 while non-students’ fell. The total barely moved. A swap, not a collapse.
November 17, 2026 at 6:55 AM
An unpopular hypothesis linking Alzheimer’s disease to the loss of crucial neurons has made a comeback.

Learn more: https://scim.ag/4zKAzY3
November 16, 2026 at 2:30 PM
To be honest, as a software engineer having to leverage AI every day to code, this is exactly the first hypothesis that crossed my mind in all these incidents. Somebody vibe-coded some change that let everything down. This maps too much with my experience.
November 16, 2026 at 8:31 PM
A physicist proposed a mathematical model in which time has three dimensions. The hypothesis aims to explain fundamental phenomena, but it has not been scientifically confirmed.

DOI: 10.1142/S2424942425500045
November 15, 2026 at 9:00 PM
That "race scientists" are wrong about race is (or should be) well known by now, but - as our Conversation piece highlights - they also get social science wrong. My co-author @jonathanrgoodman.bsky.social explains more in his recently published article

www.cambridge.org/core/journal...
November 16, 2026 at 4:00 PM
One credible hypothesis I heard is that AI agents are overloading systems
November 16, 2026 at 11:06 PM
I’m keeping the thread up to test the hypothesis that LOUD SCARY THINGS THAT MAY BE WRONG is really the path to success online. So far, yes.
November 16, 2026 at 9:06 PM
hypothesis: the mechsplo neuroablation light machine just displays very bright static, all the things seen in it are just weaponised apophenia
November 14, 2026 at 10:13 PM
fascinating hypothesis, have you tried testing it recently?
November 15, 2026 at 8:33 PM
Project Hail Mary
Wildwood
Brand New Day
The Love Hypothesis

Will replace 1: Sunrise on the Reaping

Runners up include Coyote vs. ACME and The Sheep Detectives

#letterboxd #filmsky
November 18, 2026 at 2:39 AM
It's a small but depressing number of people who want to repair the hypothesis that "It was AI" and say I "called them stupid / mocked them" in my thread for saying the 2 popular-at-the-time AI-related hypothesis seemed to be wrong. They aren't interested in even their peers saying that's wrong.
November 18, 2026 at 3:28 PM
hypothesis: airports are to Bolsonaro Jr. as hospitals are to his dad
From a friend.
November 14, 2026 at 12:37 AM
See also: the success/hype for The Love Hypothesis movie
November 17, 2026 at 8:28 PM
How about the hollow earth hypothesis?
November 18, 2026 at 3:48 PM
Yeah, we have two huge asterisks on the "proven unpopular" hypothesis and then Mayor Mamdani's continued popularity
November 17, 2026 at 6:45 AM
Achaean Smoothbrain Hypothesis
November 12, 2026 at 4:55 PM