Things I've been thinking and reading recently

From lab coats and hackathons to theorems and cruise ships

Some things I’ve been thinking about

A few weeks ago, an FT piece by Helen Pearson mentioned a trial of sturdier beer glasses to reduce violent attacks. This reminded me of another beer experiment:

Last month, I wrote about how it’s possible to build a remarkably accurate ‘AI analyst’, at least for specific types of questions. A new pre-print gave me a chance to put my theory to the test:

With hantavirus all over the news, this NYT piece by Caitlin Rivers had a useful infographic if you want to know how worried you should be (TLDR: not much at all). As it happens, we’ve previously published what is – as far as I know – the largest openly available dataset on social interactions during cruises. In short, there are a lot of them:

Some things I’ve been reading

Once, I had a book deal for half a million dollars. Now, I’m back to self-publishing by Emma Noyes

After I sign those book deals, I celebrate. I’ve done it. I’ve achieved a goal that I never thought would be achievable—and I did it twice. For an unfathomably large sum of money. It’s one of the most exciting periods of my life.

And then, the excitement wears off, and I have a full-blown panic attack.

The fall of the theorem economy by David Bessis

If you think that the hard part of a mathematician’s job is to prove theorems, let this serve as a counterexample—from the moment I conceived of Theorem 0.5, I knew it was true and that proving it would be straightforward.

The democratisation of cheating by Daisy Christodoulou

Maybe in the short-term it is easier for universities to turn a blind eye to the obvious cheating that is going on. I can see how students and professors might grumble if their traditional assessment system was changed, and perhaps students would be less likely to attend universities that had cheat-proof in-person assessments, which in turn would affect their bottom line.

Why lab coats are white by Asimov Press

Ice cream vendors, butchers, and bakers were also adopting white clothing as their professional uniforms to project the image of working under sanitary conditions. White made any breach of cleanliness immediately obvious, another way to reassure their newly cleanliness-obsessed customers of the shops’ safety and security.

Is Bluesky dying? by James Ball

There is still a huge pool of people using text-based social media daily, and Bluesky is appealing to less than 1% of them. That’s particularly disturbing given that Threads is mostly full of AI slop and out-of-date memes, and X is quite often a far-right hellhole. They’re as bad as they are and they are still winning.

Seven Phrases I Stole From Staff Engineers by Louise Deason

That disagreement is the actual information. You have just surfaced, in one sentence, the thing that was going to make the meeting fail in forty minutes anyway. Now you get to fail in two minutes instead, and reconvene with the right people.

How to apply Pixar’s “Three Pitches Rule” by David Epstein

He called it the “Three Pitches Rule.” In short, directors were asked to develop three movie ideas, rather than one. If they developed only one, they often got fixated on it and could get stuck, even if it wasn’t working well. We all often fixate on our first idea, even though it isn’t usually our best.

I Was an Enthusiastic Early Adopter of AI Scribes. Here’s Why I Stopped by Dr Benn Gooch

Because everything would be recorded and reviewed, I began to allow consultations to expand to fill whatever space the patient needed. It felt, in the moment, like better medicine. It felt like finally doing it properly. What I was actually doing was offloading the work of clinical curation — one of the most cognitively demanding and clinically important things a GP does — to the post-consultation review process.

We fund the system, not the founders by Adam Glen

Spinouts get the reviews, the policy papers, the equity debates, the HEIF funding, and the TTO headcount. Independent founders, who make up the majority of the deep tech ecosystem, get comparatively little. We have built a system to optimise the 34% while largely ignoring the 66%.

Going for gold by Polly Mackenzie

Talented people have options. Political leaders need to take responsibility for making the civil service a place that talented people want to be. You cannot simultaneously tell people the civil service is a nest of incompetents and expect the competent to queue up to join.

How to fundraise one million million dollars, part 1 by Natalie Cargill

Being good at fundraising, in the popular understanding, means being good at schmoozing, the “charm offensive”, and other such social manoeuvring. “Being good at” is also something of a misnomer, as you’re either blessed with the ineffable social magic or you’re not.

Being “bad at fundraising”, by contrast, means one is reassuringly inept at social wizardry, and — by implication —good at Real Work (such as computer, research, and computer). One is bad at vibes, but good at skill.

Can AI Make Better Decisions Than Pharma Executives? I Tested It on the Biggest Call in Oncology History (on a Time Machine) by Liang Chang

It found the consensus position: the position that the evidence, taken as a whole, most strongly supported.

The problem is: the right answer in this case was a contrarian bet. Merck’s decision to aggressively enrich for PD-L1 >=50% went against the weight of available evidence.

Purity Tests and Pennies by Jess Steier, DrPH

The next time someone at a podium tells you science communication has never been more important, ask them what their organization is doing about it. Not in the abstract. In the budget. In the job postings. In the contracts they signed this quarter. That is the question. Everything else is applause.

Trust and discretion by Daniel Greco

Faculty members face various incentives to inflate their grades—most obviously, it makes students happy, which helps with course evaluations—and once other faculty are inflating grades, few faculty are willing to risk scaring students away by developing reputations for harsh grading.

How to win hackathons by John Kim

The best way to win hackathons is recognizing what a hackathon actually is. In my opinion, a hackathon is just marketing. The sponsors are paying to be there. They want developers using their tools, creating demos, generating content that makes them look good. The organizers want a story. If you understand that, you can work backwards from it.