Quasi-Riemann, Biologic Discovery, and Swanson's "Undiscovered Public Knowledge"

In recent weeks, we've seen AI make impactful discoveries like the Quasi-Riemann proof and biologic discoveries aiming to cure diseases.

The common theme among these discoveries: A model's ability to make connections between disciplines, using existing knowledge.

From the Quanta article, "AI Proved the Quasi-Riemann Hypothesis. What Happens Now?":

"Moreover, contrary to what mathematicians had long expected, the proof didn’t need to introduce any new concepts — it instead unearthed hidden connections between old ones. “The proof is a surprising synthesis of well-known ideas,” said Alexander Dunn(opens a new tab) of the Georgia Institute of Technology."

And from the OpenAI article, "Using thousands of AI agents to uncover new leads in biological data":

"But understanding what these results mean for human disease requires connecting complex experimental findings with decades of scattered scientific knowledge. ... New discoveries often emerge when researchers from different fields bring their expertise together and connect familiar ideas in unexpected ways. AI can significantly scale these connections by linking new data with knowledge scattered across disciplines."

"Sometimes the clues are already there, scattered across different fields—we just haven’t connected them."

In the past, I have thought: Humans have short lifespans and limited brain capacity. What if you were able to hold all of mankind's academic knowledge in your head at once? Limited not to a single perspective from your chosen specialty, but with the ability to look at a problem from every possible side at once.

https://x.com/GavinRayDev/status/2054963487139574000

Rarely is a thought original, and in this case, I am far from the first. In the 1980's, a researcher named Don Swanson published a paper on what he called "Undiscovered Public Knowledge". Rather than try to poorly paraphrase, the LLM explanation does a fairly succinct job:

Swanson used this framework to make two impactful connections: Proposing that fish oil would treat Reynaud's syndrome, and that Magnesium deficiency may cause migraines.

Both of these were cases of "FOO treats BAR" and "QUX is worsened by BAR", with the missing connection that FOO ought to help QUX.

I think we're going to see an exponential increase in discoveries of this sort as AI's resources and abilities grow. And that's beyond exciting.

AI (Dis)use Disclaimer: No part of the prose was machine-generated. You will not find machine-written prose on this blog. I consider it deeply disrespectful.

OpenAI Codex was used to generate the OpenGraph image for this post