Intro

There is a shift happening in the complex world of proof. Simulation and probabilistic approaches are increasingly accepted as ‘good enough’ in areas traditionally dominated by exact proofs. Persuasion depends on the degree of certainty needed.

Adam Kucharski is a professor at the London School of Hygiene and Tropical Medicine, and also the author of three books, Proof: The Art and Science of Certainty, The Rules of Contagion: Why Things Spread--And Why They Stop, and The Perfect Bet: How Science and Math Are Taking the Luck Out of Gambling.

Greg and Adam discuss the versatile concept of 'proof', examining how it applies differently across mathematics, law, medicine, and practical decision-making. Adam discusses the challenges of proving concepts under uncertainty, particularly during the COVID-19 pandemic, and the role of intuition versus formal modeling in various fields. They also explore the crossover of epidemiological principles into finance, marketing, cybersecurity, and online content dynamics, illustrating the universal relevance of contagion theories. 

The episode highlights how simulation and probabilistic approaches are increasingly accepted in areas traditionally dominated by exact proofs.

*unSILOed Podcast is produced by University FM.*

Episode Quotes:

The gap between science and policy

09:25: One of the challenges we had in COVID is this dimension of a problem where all directions had a lot of enormous downsides, and countries were having to make that under pressure. And even one of the things that I think I did not really appreciate at the time was, even later in the year, when a lot of these questions about the severity, a lot of these questions about transmission, had really been resolved because we had much better data. We still had a lot of this tension demanding, "Oh, we cannot be sure about something," or "You know, we need much, much higher evidence." And I think that is the gap between where kind of science lies and where policy lies.

It’s not the content, it’s the contagion

37:59: I think a lot of people think about the content, but obviously it is not just, "It is something goes viral." It is not just about the content. It is not about what you have written; it is about the network through which it is spreading. It is about the susceptibility of that network. It is about the medium you use. Do you have it that lingers somewhere? Is it just something you stick on the feed and it kind of vanishes? So, there is a direct analogy there with the different elements and how they trade off in ultimately what you see in terms of spread.

What human networks can’t teach us about machines

46:35: One thing that is really interesting about computer systems is the variation in contacts you see in the network is enormous. You basically get some hubs that are just connected to a huge number of computers, and some are connected to very few at all. So that makes the transmission much burster.

It is not like—so humans have some variation in their contacts—but most people have about 10 contacts a day, in terms of conversations or people they exchange words with. Some more, some less, but you do not have people generally have like 10,000 contacts in a day, whereas in computers you can have that. So it makes the potential for some things to actually persist at quite low levels for quite a long time because it will kind of hit this application and then simmer along, and then hit another one and simmer along.

Show Links:

Recommended Resources:

• Euclid
• George E. P. Box
• William Sealy Gosset
• P-value
• Ronald Ross
• Jonah Peretti
• Duncan J. Watts
• Amazon Web Services
• Monty Hall

Guest Profile:

• AdamKucharski.io
• Faculty Profile at London School of Hygiene & Tropical Medicine
• Social Profile on BlueSky

Guest Work:

• Amazon Author Page
• Proof: The Art and Science of Certainty
• The Rules of Contagion: Why Things Spread--And Why They Stop
• The Perfect Bet: How Science and Math Are Taking the Luck Out of Gambling
• Substack Newsletter
• Google Scholar Page
• TED Talks