Adam Kucharski

Beyond Belief The remarkable story of the global movement championing the idea that evidence, not opinions, should guide our decisions

📢 My book, Beyond Belief: How Evidence Shows What Really Works, is out 28th April. 🚨🚨

“an incredibly engaging popular science book that illustrates why we should care about good quality evidence.”

Preorders *really* help! 🙏
@princetonupress.bsky.social
press.princeton.edu/books/hardco...

“I got AI agents to rewrite my company docs in the voice of a 1990s football commentator, it did a great job”

We see these kinds of fun trivia examples all the time. But I’d much prefer to see how well they did at reproducing your last tax return. Or that thorny data analysis you spent ages on.

"A development programme for non-fiction writing on health and being human"

£2,000 bursary
Sessions with publishing team, library research specialists and editorial team
Mentoring by a published author & a non-fiction editor
Meeting with an agent
Travel and Access fund
Etc

This is a really fun dataset to play with - thanks to @adamjkucharski.bsky.social for creating it! I'm looking forward to seeing some #TidyTuesday visualisations of it! 📊

And the quiz is still open if you'd like to add yourself to the original data: probability.kucharski.io

Wellcome Collection Non-Fiction Awards | London Writers Centre Learn more about Wellcome Collection Non-Fiction Awards, a Writer development project from the London Writers Centre.

Are you a UK-based writer from an underrepresented group? The Wellcome Collection Non-Fiction Awards are now open for applications: www.londonwriterscentre.org.uk/project/well...

Got a bit delayed, but sending this week!

People bragging about their .md files...

Are we entering the era of Understanding-as-a-Service? kucharski.substack.com/p/how-much-t...

Does ~10% of a required task. Thinks it’s a real achievement.

AI agents are low agency.

Have you heard about the mathematician who published her first paper in Nature at the age of 10?

Yep, "it's a great tool as long as you already have a good idea of the answer" isn't ideal...

Nope – just scattered across reports and websites

New post, on whether I could get Claude Code to complete a data task that had taken me AGES a decade ago…

kucharski.substack.com/p/how-much-t...

Proof - Chalkdust We review the second of this year’s nominees for Book of the Year

"This book does a brilliant job of showing that mathematics is about more than we learn at school."

Great to see @chalkdustmag.bsky.social have shortlisted Proof for one of their books of 2025: chalkdustmagazine.com/book-of-the-...

Crises can be forgetten quickly – especially health ones, where the stories about the invisible enemy soon get blurred.

After Covid by @jasongale.bsky.social is an important account of how things actually played out, full of details that were crucial then, and will be crucial for next time.

Can’t believe I’ve never seen this great Collatz conjecture visualisation before…

As Paul Erdős once put it: "Mathematics may not be ready for such problems”

Image by github.com/dkobak/colla...

1908: the Lancet, one of the most respected scientific journals, calls for 18 age limit on reading in bed amidst a moral panic surrounding children becoming "addicted" to novels, which were "designed to keep kids hooked" and destroy their attention/mental health

Lots of headlines use phrases like 'realistic possibility', but how do people interpret this?

New post, with insights from over 150,000 human judgements about probability-based phrases: kucharski.substack.com/p/insights-f...

Fascinating read on Cantor and Dedekind: www.quantamagazine.org/the-man-who-...

So basically two pub bores from Kent are annoyed a woman from Manchester didn't let them win an election in her back yard?

The distribution is interesting. Clearly the English language has a dearth of ways to refer to an event that has a roughly 30%–35% chance of happening: too much for “unlikely”, too precise for “could happen”, too low for “about even”. We need to invent a new word for this!

The outputs are based on an online quiz I’ve been running over the past few weeks, with 5000+ participants and counting. It seems to have really resonated with people, so thanks for everyone who’s taken part – hopefully there’s lots in the resulting data for people to explore.

Introducing CAPphrase (Comparative and Absolute Probability phrase dataset), an open access dataset containing over 150,000 probability-based language judgements: adamkucharski.github.io/CAPphrase/

There are some nice examples out there, but largest public dataset (at least that I'm aware of) is ~120 responses, so this is a much larger sample size, and also includes more phrases plus comparative vs absolute estimates.

Fair point. Have seen quite a few "we're going to do an extremely complex analysis and develop software libraries with 3 months of postdoc time" in grants over the years...

I've also noticed "surely AI can just do the data analysis on the cheap?" creeping into grant reviews.

This creates a big risk that researchers will collect valuable data – then throw away most of that value with a conceptually flawed or misrepresented AI analysis.

I think learning the basics is still crucial. AI tools can make it much easier to switch coding languages or understand how to edit models (we've successfully used in teaching for this), but there will be a lot of bad outputs if people don't understand what's happening conceptually.

How likely is ‘almost certainly’? The scourge of weasel words Phrases to describe probability are getting lost in translation — and have helped to cause at least one military catastrophe

I’m in The Times today talking about how we judge probability-based language and what happens when words mean different things to different people.

This follows an online quiz I’ve been running at probability.kucharski.io over the past few weeks, with 5000+ participants and counting.

Why the search for proof can’t be separated from faith | Psyche Ideas Rather than being an enemy of empiricism, belief in what can’t be known is part of how we gain knowledge, even now

This refreshing Idea by the mathematician Adam Kucharski argues belief ‘beyond the limits of reason’ is at the heart of science. From early scientists’ firmly held religious beliefs to mathematical proofs, faith and reason are not opposites, but deeply intertwined

"Excellent suggestion. Would you like me to implement that?"