Things I've been reading recently

From Crick and Claude to wine and writing

Francis Crick and Nature by Matthew Cobb

Perhaps more significant than what Crick said in that article is how Nature responded to it. Crick submitted the article on 1 January 1970; the journal received the manuscript the following day and accepted it within hours, with no changes. It went straight to production and appeared at the end of January.

How Hard Can Quant Trading Really Be? I Tried It to Find Out by Lauren Leek

The difference between me, my Python scripts, and a hedge-fund quant isn’t insight - it’s scale. They have cleaner data, faster pipelines, and infrastructure that turns risk into opportunity.

How to write a good rubric (for humans and AI) by Daisy Christodoulou

If you push these strategies to their ultimate limit they become entirely self-defeating. You end up with football teams trying not to play football and writing lessons that are about avoiding writing. Ultimately, if you want to win football games you have to try and play some football. If you want to be a good writer you have to write something. Neither writing nor football are exercises in trying not to make mistakes.

I'd like to thank the Academy by Oliver Johnson

At some stage when I joined Bristol, I was doing a tutorial about changing the basis of a matrix. It’s a kind of recipe to follow: you calculate some things, you put them in a table, you invert the table, you figure it out, some numbers come out in the end. But I remember asking the students “why are you doing this?” and they didn’t really know. It was just a problem on a problem sheet, a topic in a lecture course.

On research careers in academia and industry by Andrew Lampinen

In both academia and industry, a portion of one’s job is to advocate for one’s work.

In academia, this includes writing grants, creating packages for tenure or promotion, etc.

In industry, this similarly takes the form of writing packages for promotion, and advocating for why your work should get resources. Many large-scale industry projects started as small ones that were given more resources because they demonstrated some preliminary success — much like the pilot experiments one might describe to convince grant reviews.

Overfitting to theories of overfitting by Ben Recht

But let us not, um, overfit, to double descent. The thing to take away from double descent is that you can see it. But you might not see it. Depending on how you define model complexity, you can see all sorts of things.

Pain by Polly Mackenzie

There is broad agreement that more people should be able to work, participate and live with dignity. But political energy has increasingly flowed into coercion rather than support: tougher assessments, harsher conditionality, and suspicion in place of trust. This persists despite weak evidence that punishment improves outcomes, particularly where pain and fluctuating conditions are concerned.

Forget learning a new language. Instead, focus on learning how to communicate by Souvenirs

Hard is when you actually try to learn the language, and Easy is when you just say things more loudly and slowly, expecting people to eventually understand.

Defence Isn’t Optional. We’re Just Pretending It Is. by Martha Lane Fox

Take a whole-of-society view seriously and the tech stack stops being a procurement issue and becomes a sovereignty issue. Who owns the cloud we rely on? Where are the data centres? Who designs the chips that power AI models and weapons systems? Digital transformation looks very different once you imagine those systems under stress or attack.

claude code psychosis by Jasmine Sun

The problem is not that your friend is horribly uncreative. It’s that most people’s problems are not software-shaped, and most won’t notice even when they are.

How to be a wine Midwit by Kristian Niemietz

Cremieux shows that if you serve a wine judge the same wine three times without telling them that they’ve rated it before, there’s a good chance that you’ll get a different rating every time.

Why Can’t I Enjoy Anything Anymore by John Kim

My dad wasn’t broken because he didn’t have hobbies. He found his joy in building, same as I have. He just never apologized for it or tried to force himself back into a mold that didn’t fit anymore.

The question isn’t “How do I enjoy games again?” The question is “Am I okay with being someone whose joy comes from building?”

AI’s exploration deficit by Azeem Azhar & Nathan Warren

AI pushes work toward data‑rich problems, and some foundational questions where data is sparse, go unexplored. We face an exploration deficit where AI will do well at exploiting what we already know, but it is eroding the incentive to discover what we don’t.

Toys with the highest play-time and lowest clean-up-time by Joanna Bregan

Apparently every toy that I find easiest to clean up has magnets on it. Maybe I feel the satisfaction of clicking them together as I clean them up. Cleaning becomes a little like playing.

The 80% Problem in Agentic Coding by Addy Osmani

Given free rein, agents can overcomplicate relentlessly. They’ll scaffold 1,000 lines where 100 would suffice, creating elaborate class hierarchies where a function would do.

---

Cover image: Olena Bohovyk