Will Lowe
banner
conjugateprior.org
Will Lowe
@conjugateprior.org

Señor Research Scientist, NPC at the Hertie School in Berlin 🇩🇪 via Princeton, Mannheim, Edinburgh and a bunch of other ivory towers that will probably be billiard balls and decorative boxes by the end of the decade.

Rome Statute appreciator. .. more

Computer science 32%
Political science 28%
Pinned
For Monty Hall problem aficionados: a #causal DAG, with explanation in the alt text.

Who knew that the M in M-bias stood for Monty?

One day, by a very unlikely combination of circumstances, he actually reads one of these books. It lands quite hard. The persona pauses, people wonder what's up, there's some public flailing, and it melts into a weirder but more plausible 15yo (who is bundled away from public view by handlers.)

The Appointment with AI in Samarra

Ha. I nearly linked to it but ran out of space.

Yes, that's the contrast of interest

The relevant observations are imo 1. MH is a causal inference problem first and a probability problem second, and 2. the association usually called collider bias is information you can make use of for decision rather than just being, as it normally is, a problem

I mean, they _could_ (le yikes) but it seems like there are stacks of potential natural experiments using everything from immigrant populations to high school curriculum shocks.

The slow but happy convergence of experiments & design inference and modeling & simulation strategy continues.

Apropos, I'm beginning to see why @stephensenn.bsky.social is so enthusiastic about John Nelder (general balance & design inference) and also John Nelder (GLMs & hierarchical models).
Very interesting perspective! I often have the opposite thought: "Experimenting like a modeller"
Perhaps the two could co-exist for better workflows. Extensive exploration on simulated data should be the norm before running experiments to make sure the design is capable of the intended inference.
"Validate With Simulated Truth: A first habit is to test whether an analytical pipeline can recover known conditions."

Very good advice below. So much COVID nonsense (e.g. 'immunological dark matter') basically came down to a non-identifiable model that hadn't been properly tested.

Reposted by Will Lowe

Very interesting perspective! I often have the opposite thought: "Experimenting like a modeller"
Perhaps the two could co-exist for better workflows. Extensive exploration on simulated data should be the norm before running experiments to make sure the design is capable of the intended inference.
"Validate With Simulated Truth: A first habit is to test whether an analytical pipeline can recover known conditions."

Very good advice below. So much COVID nonsense (e.g. 'immunological dark matter') basically came down to a non-identifiable model that hadn't been properly tested.
Modelling Like an Experimentalist
Dahlin et al. (2024) apply experimental thinking to a model of mosquito-borne disease transmissions.
onlinelibrary.wiley.com

Sharpies: Because sometimes you need forecast *regions*

Reposted by Will Jennings

You can wear a suit, you can be in a complicated relationship with a suit, or the suit can wear you.

Bet you didn't know that flags can have familiars.

A real serial monogamist. Lots of issues tbh.

"Men will literally suppress Catholicism in England rather than go to Mass" (Thomas More, probably)

Ah, yes. One of those parts of Berlin that would be improved by converting it into almost anything else. A multistory parking garage, a water park, an even bigger Bud Spencer museum. Anything, really.

Antinous: "Come on Odysseus, put it away. You know you're not going to use it."
Eurymachus: "The bow he means!"

Also, given what I see being pulled out of the Spree on the regular, I think it might be full.

Fun fact: there are more illegally parked cars in the photo than there are scooters. Obviously I've redacted them because this is Germany and cars are very sensitive and private things that don't take criticism well and must be allowed to be able to park anywhere for free.

See, this is what happens if you feed them

🤬

I just learned that we could have had country-wide fiber optic in the early 1980s but Kohl cancelled the project because copper was cheaper.

(Flag signals stopped being accepted in the late 1960s ofc)

😢

Nobody was sending spam at 1 euro per character, especially [ahem] in German.
Germany's telegram era ends with final rush of thousands
BERLIN (AP) — The era of the telegram came to an end in Germany with a final flourish of thousands of the once-popular message service, the country's main postal service, Deutsche Post, said Wednesday...
apnews.com

So, on that last question...
[throws open doors to reveal a vast multistorey library]
First floor: history
Second floor: couches
Third through tenth floor: psychopathology

It's the perfect word because it's also almost a pictogram

_⧹\Igel/⧸_

1 Minute zum Kaufen
6 Tage bis zur Lieferung
9 Monate bis zur Genehmigung durch die Eigentümergemeinschaft

Wir können alles. Außer Hochgeschwindigkeit
Solar’s price drop is astonishing: panels are now 98% cheaper than when I first analyzed them in 2004.

Today, building a fence with solar can be cheaper than using wood.
Solar’s price drop is astonishing: panels are now 98% cheaper than when I first analyzed them in 2004.

Today, building a fence with solar can be cheaper than using wood.

You might like stationary bandit theories of state formation.

[very Oxford 1960 voice]
so the question *tamps pipe* is whether 'noumena' can be coherently understood as plural.

Nick's free

Es gibt eine Schlange, und wir stehen drin.
Be clear, the American far right are coming for British democracy.
Be clear, the American far right are coming for British democracy.