Tom

New 2026 US House & Senate election forecasts, using available political science knowledge & our methods from “If a Statistical Model Predicts That Common Events Should Occur Only Once in 10,000 Elections, Maybe It’s the Wrong Model” with @DannyCEbanks and @Jonathan_N_Katz. GaryKing.org/electit

One of the more interesting parts of this famous chapter for me is this is a fairly BAD way of thinking about surgery. The stereotypical rockstar surgeon with total hierarchical control is directly challenged by all kinds of research on team outcomes in ORs including v famously psych safety research

Making myself angry for no reason at all remembering the American transit planner that went to Amsterdam and was appalled at how little stakeholder engagement they did in favor of actually building shit.

Also an amazing watch for #F1 fans

wild that this is exactly what using a computer is like now

What is not scarce has no value

I wish there were a hundred more Cat Hickses at every level of this tech doing work like this (and kudos to whoever *is* already doing that)

Kimi win from the back, again?

Wait If submarines don’t swim, what do they do Surely the don’t sail

I have often said people are essentially arguing that engineering is impossible when they say there’s no way to build a reliable system out of unreliable parts. it‘s the most ordinary thing in the world

I want to expand on this in a short thread. Fascism sucks. It is also not going away: it is an effectively permanent part of modern representative politics because it is an effectively permanent part of human nature. We have to be prepared to push back against it, basically forever.

If the finding is that AI is bad or biased at something, you need to be much more careful. You cannot claim that because GPT-4 fails at something, that AI is bad at that task, because current or future models may do it. So what can you claim? These are just some examples: 3/

Holy shit Leclerc/Piastri! Amazing racing.

What happened to owning the means of production?

Spreadsheet work is a large part of the economy and science too.

The beaver urge to dam running water

Finally. We need more actually European banks. www.euronews.com/business/202...

Fable speaks Darmok and Jalad at Tanagra

turn on the sound TURN ON THE SOUND

This is one of those arguments that sounds like slopulism but is actually, if anything, understated. 56% of Colorado river water goes to GROWING FEED FOR LIVESTOCK IN THE DESERT.

This is the French nuclear deterrence doctrine

I’m convinced, upon rereading, that this simply misunderstands how either or both of how LLMs or watermarking works daringfireball.net/2026/08/anth...

Email + atproto anyone?

umsteigen.app · Berlin 🚇 14 Aug 404 / 500 · Daily commuter 🟢–🟡–🟢–🟡–🟢

⚪️ Відбій загрози балістики.

#F1 pundits: make zero effort to understand energy recovery and deployment #F1 pundits trying to explain the difference in pace between George and Kimi:

This what pivot_wider() was born for

It’s not intelligence. It’s what you do with it.

I’ve learned to code at 13 on a Psion Organiser II. That thing could display 2 lines of 16 characters. The code needed to be in my head. Still one of the most interesting summers of my life.

Any tool can be very sloppily used. You’re responsible for what you put out in the world no matters the tools you’ve used to produce it.

Reverse sophon times

Continuous learning is not gonna be solved in the enterprise. It’s gonna be solved for SMEs.

“What are you gonna do? Stab me?”

Radicalized against AI

This is a wonderful app for discovering what’s around you. I love that the Wikipedia projects makes apps like this possible. apps.apple.com/gb/app/inqui...