Why aren't there more AlphaFolds? | Nima Keivan

This essay began as a talk at the 2026 Gold Lab Symposium. You can watch the talk itself here. The content is the same, but I’ve expanded on it more here.

The universe doesn’t give up its secrets easily. Every time we learn some truth,<br>we depend on prior, equally hard-won knowledge, and usually some luck. The easy<br>truths are getting rarer,<br>so the excitement around speeding up science with deep learning models is<br>understandable, especially after LLMs surprised us by learning so much just from<br>predicting text. Today’s frontier LLMs are trained on almost the entirety of<br>human scientific output. So why haven’t they transformed our understanding of the<br>universe? And why is AlphaFold, a deep learning model for predicting protein<br>structures (and very much not an LLM), one of only a handful of models to have<br>transformed science? Why aren’t there more AlphaFolds?

The answer hinges on the difference between training models on human-generated<br>text, which is itself downstream of knowledge acquired the hard way, and<br>training them on observations at the limits of science where the definitive text<br>is yet to be written. It’s easiest to see this by studying the process of<br>scientific discovery. So let’s start by looking at two examples, with the first<br>being the story of one of the most deadly diseases of the age of discovery:<br>scurvy.

Scorbutic and Confused

On July 8th 1497, the Portuguese mariner Vasco da Gama sails from Lisbon to<br>find a path from Europe to India via sea, commanding an armada of four ships and<br>170 crew. Six months later, in January 1498, after sailing deep into the South<br>Atlantic Ocean in search of westerly winds, they arrive at the Mozambique coast<br>near Quelimane. The ships’ journal1:

“Many of our men fell ill here, their feet and hands swelling, and their gums<br>growing over their teeth, so that they could not eat.”

Vasco da Gama's path to India around Africa. The stop at Mozambique is circled in red.<br>We know what’s happening to these sailors. Their vitamin C reserves are depleted<br>after six months at sea eating salted meat and dried crackers. You need vitamin<br>C (amongst other things) to keep collagen intact, which is needed to keep your<br>gums attached to your jaw. Without it, you get scurvy.

Three months later, the armada sails into Malindi (modern-day Kenya). Local<br>trading craft approach the ships. Here’s the journal again:

One was laden with fine oranges, better than those of Portugal

Five days later, the sick miraculously recover:

“… on arriving at this city all our sick recovered their health, for the<br>climate (“air”) of this place is very good…”

It was the “climate”, not the tasty oranges the crew ate right before they got<br>better. To be fair to them, all the novelty they encounter sailing into Malindi<br>confounds exactly what cured them. But as we’ll see later, attributing malady<br>and recovery to the climate isn’t uncommon for the era.

The armada keeps going, crossing the Arabian Sea and reaching India, becoming<br>the first Europeans to do so by sea. The same crossing on the way back takes<br>three months against the monsoon. Half the remaining crew die. The rest limp<br>back to Malindi, each ship having only seven or eight men left fit for duty. The<br>journal again:

“The captain-major sent a man on shore with these messengers with instructions<br>to bring off a supply of oranges, which were much desired by our sick.”

Despite not knowing anything about vitamins or collagen, the crew remembered the<br>oranges, asking for them explicitly. By the time the crew returns to Lisbon,<br>half the original crew of 170 are dead, many from scurvy.

Subscribe for occasional human-written essays on machine learning, robotics, startups, and more.<br>Subscribe<br>Two hundred and fifty years later, in 1747, the Scottish physician James Lind<br>runs one of the first controlled clinical trials2. On the HMS Salisbury, Lind<br>gives twelve sailors with scurvy six different treatments, including oranges and<br>lemons. Unsurprisingly, the citrus do really well. Lind writes in A treatise of<br>the scurvy, in three parts:

Lind on the treatment that gave the best results in the trial<br>What does Lind conclude after the experiment? That scurvy is caused (amongst<br>other things) by the “want of fresh vegetables and greens” and by the “moist sea<br>air … rendered still more noxious, by being confined in a ship without due<br>circulation”. Remember this diagnosis? Da Gama’s crew came to the same<br>conclusion in Malindi. Lind’s subsequent advice to the navy to keep ships<br>ventilated and provide fresh vegetables does nothing to prevent scurvy.

Lind on the noxious sea air causing scurvyLind on the lack of fresh vegetables causing scurvy<br>And what of the oranges and lemons that had such a positive effect? Lind<br>instructs to preserve them on long journeys (before the invention of<br>refrigeration) by boiling them into a “rob”:

Lind on preserving citrus on long voyage as a boiled "rob"<br>What Lind didn’t know was that boiling the...