Philip Ball

Everyone in my lab plays with AlphaGenome. It is good at what it does, but what it does is very limited, because it is only trained on data from a handful of cell lines. To me the big gap is in the complexity of the human body, in all the cell types and how they change over time in development – and all that data is missing. We have got pretty good information on what sequences are important, but when we went into AlphaGenome it couldn’t predict any of them because it has never seen data from a developing limb in an embryo or a developing eye. It is good at what it’s trained on, but not very good at the complexity of cell types in the body.

And here is an excerpt from the transcript of my interview with someone who works on human gene regulation. Says it all.

Dreams of a synthetic genome In June, 2025, the Wellcome Trust announced an ambitious £10 million UK project called the Synthetic Human Genome Project (SynHG) and claimed it “will unlock a deeper understanding of life, leading to...

This might be relevant to the discussion too.
www.thelancet.com/journals/lan...

I didn't - thanks so much Ewen, looks really interesting.

What distinguishes the elephant from E. coli: Causal spreading and the biological principles of metazoan complexity - PubMed Jacques Monod famously said that 'What is true for <i>E. coli</i> is true for the elephant.' While this might be correct in the basic sense that both use nucleic acids and proteins, it is no longer clear that they use them in quite the same way. The many qualitative differences in the biomolecular c …

You got there just before I added this to the thread. Yes indeed!
pubmed.ncbi.nlm.nih.gov/37194562/

I'll say more about this in a forthcoming article.
Also, FWIW: I don't consider synthesizing a genome based on existing genomes and putting it in a living cell to be making "synthetic life". That sets the bar way too low.

What distinguishes the elephant from E. coli: Causal spreading and the biological principles of metazoan complexity - PubMed Jacques Monod famously said that 'What is true for <i>E. coli</i> is true for the elephant.' While this might be correct in the basic sense that both use nucleic acids and proteins, it is no longer clear that they use them in quite the same way. The many qualitative differences in the biomolecular c …

Over the past week I've had discussions with several experts who *really* know their stuff in this regard, and they confirmed and deepened my scepticism. It's important we don't make casual conceptual leaps from bacteria to humans. That has misled us in the past.
pubmed.ncbi.nlm.nih.gov/37194562/

AlphaGenome: On the Promise and Limits of AI in Science PHILIP BALL | According to the old story, we once thought that all this other DNA was just “junk,” a term coined in this context in the 1970s. It was accumulated over the course of evolution, for exam...

For that, qualitatively different considerations come into play. Metazoan genomes are in many ways *not like* prokaryotic ones. It's not just more of the same. Coping with different cell types is a massive issue. I discussed some of these things here.
www.marginaliareviewofbooks.com/post/alpha-g...

AI can write genomes — how long until it creates synthetic life? The Evo2 genomic language model can generate short genome sequences, but scientists say further advances are needed to write genomes that will work inside living cells.

Ewen's writeup of the new Evo 2 paper in Nature is reliably good. I can imagine this foundation model will be useful for doing synthetic biology in microbes.
But it's important to recognize that making this work for eg humans is not just a scaling-up issue...
www.nature.com/articles/d41...

Was the Turing Test always meaningless? A meeting to celebrate the 75th anniversary of Alan Turing’s famous paper drew a fairly unanimous judgment from the expert speakers

Thanks Eddy. There's a debate to be had, for sure. It's an exaggeration to say that *no one* takes the TT seriously, but I stand by the claim that few experts do. That was certainly the consensus of this recent Royal Society meeting:
www.thenewworld.co.uk/philip-ball-...

I never knew until now that Electric Avenue in Brixton was called that even in the late 19th C. It was the first street market to be lit with electric lights. South Londoners can't help but now associate it with Eddy Grant. In the 1910s Madame Ziska told fortunes there with a crystal ball.

Thought Machine (d'Odiardi) - Kook Science

TIL about Savary d'Odiardo's Thought Machine and its use by the psychic Cheiro. Fabulous stuff. I wonder if it still exists?
hatch.kookscience.com/wiki/Thought...

There are qualitatively different challenges for eukaryotes and especially metazoans. I have been speaking to some expert folks about this just in the past week or so, and they confirmed all of my feelings about limitations, and added to them. But as a tool for basic synth bio it should be useful.

I've just read Ewen's news story, not yet the paper. The news story offers some appropriate caveats, but it seems nice work and all quite plausible. And... using this for more than microbes, and especially for humans, is of course not just a scaling-up issue...

🙏

"Do not make vermilion or lead white or orpiment outside a professional setting" would be good advice though.

I've come across folks who do the mineral-grinding thing, and love the idea. And also of course lots of people making plant dyes. But "synthesizing pigments" implied something even beyond that - real (and potentially dangerous) chemistry.

Oh goodness Sigrid, that is the best kind of feedback I could wish for regarding Bright Earth! Thank you!
(And yes, I had that same thought. She could of course find Theophilus's recipe for vermilion in my book, but I'd not recommend trying it...)

Rendered in Handmade Pigments, Rupy C. Tut's Warriors March Toward Belonging For Rupy C. Tut, making her own pigments and painting warrior-like figures is an act of devotion and rebellion.

I like both the sound and the look of this. But goodness, can you really talk about "handmade pigments" and not explain what is meant by that? Does Tut synthesize these compounds from scratch? I'd love it if so! They're certainly gorgeous colours.
www.thisiscolossal.com/2026/03/rupy...

You mean we have enough problems already?

Time to dig up the future In a terrible irony, the green technologies we need to move away from fossil fuels rely on minerals that can only be extracted with terrible environmental consequences

In my latest column for The New World I take a look at the geopolitics of rare earth elements.
www.thenewworld.co.uk/philip-ball-...

One possibility, though probably not a good one, is simply not to report things he says, as they have little meaning. The problem with that is that evidently his minions are acting on the things he says - so in that sense they do have meaning, or at least consequences.

I don't know how much thought news agencies are giving to how they might deal with the fact that Trump is deranged and has child-like mental functioning (forgive me kids). There's not really a precedent and it's not obvious how to handle it. But pretending it away is not an option.

Galileo’s handwritten notes found in ancient astronomy text Discovery sheds new light on how famed astronomer came to lead a scientific revolution

Looks like a fascinating insight into the young Galileo - I imagine historian Owen "The Great Copernicus Chase" Gingerich would have loved this marginal discovery.

Galileo’s handwritten notes found in ancient astronomy text | Science | AAAS www.science.org/content/arti...

The quest to understand where atoms end Atomic size measurements like van der Waals and covalent radii are central to chemistry, but are they grounded in reality?

Where is the edge of an atom? My article for @chemistryworld.com on efforts to figure out the size of atoms, and why it matters.
www.chemistryworld.com/news/the-que...

Alchemy: A Curated Story of Transmutation This article offers a detailed review of Philip Ball’s recent book on alchemy. It situates and interprets the work within contemporary debates regarding the science-pseudoscience distinction, the h...

A wonderfully in-depth review and critique of my book Alchemy.

"Philip Ball’s Alchemy is an erudite and sumptuous work that aims to place alchemy as one of the great intellectual adventures in the history of ideas."
www.tandfonline.com/doi/full/10....

The murals of the Villa Arianna in Pompeii.

One for pigment junkies like me: an analysis of the colours on the murals of the Villa Arianna in Pompeii.
onlinelibrary.wiley.com/doi/full/10....

😱

Can we extract the Reform nonsense from Labour? These immigration proposals are obscene, but the government can be forced to think again.

No administration with a moral core would have put these plans forward, let alone try to enact them without a vote. Once these proposals are in place, they will not be reversed. Make them think again iandunt.substack.com/p/can-we-ext...

The quest to understand where atoms end Atomic size measurements like van der Waals and covalent radii are central to chemistry, but are they grounded in reality?

How big are atoms? That depends where you decide to put the edge. My article for @chemistryworld.com on the latest efforts to define atomic radii.
www.chemistryworld.com/news/the-que...

😱But also ✊🎉