Tom Chivers... Bayes Theorem All The Way Down

Is Everything Predictable?

Author

Curious Worldview
July 09, 2024

hey mate! Just in case you’ve been forwarded this email but would like to subscribe - click the link here.

Is Everything Predictable?

Other episode of the podcast that suit this episode...

Tom Chivers is a prolific science writer whose written for Buzzfeed, The Telegraph, Unherd, published books, written for loads of other publications as well and now writes for Semafor’s daily flagship email (something I read everyday)…

But here Tom is today to discuss his book about Bayes and how Bayes’ remarkable theorem explains the world. It’s is a book about Bayes Throerom which to put it simply… is an equation to calculate probability.

Now, my Talebian listeners will recognise a contradiction to our worldview in the title here… everything is predictable? how often has Taleb’s quotes, how can we predict a future of infinite possibilities based off a finite experience of the past appeared on this podcast? We get into Chivers differences with that Talebian worldview, but as well, there is top to bottom what is Bayes theorem, why does it matter, the role of this theorem at the foundation of all of these LLM’s and therefore much of AI. a neat little anecdote of Chivers family member, Sir John Maynard Keynes and plenty more as well!

Forward this email or share this podcast episode with a fellow Talebian or probability nerd…

Here is a transcript of the opening exchange from the conversation…

Ryan
Okay. John Maynard Keynes was your great great uncle.

Tom Chivers
Yes, yes, that is right. Yeah. It's a bit of a name drop in the middle of the book that I wasn't sure whether to include it, but it is quite cool. So I like to mention it from time to time.

Ryan
In his, his discography of writing, did he ever talk directly to bayes?

Tom Chivers
You know what he did? He obviously didn't talk directly in the sense of actually... Yeah, yeah, yeah. But his... Now, I didn't go into a great deal of detail about it in the book, and I'd like to pretend that's for a better reason that I didn't fully understand it, but the... Keynes wrote what I think was a kind of defense of Bayes, I'm somewhat someone might write in and correct me if I've got this exactly backwards. But I think Keynes wrote to think one of the problems with Bayes is finding out where you get your priors from, which I'd say we'll get to get into at some point. But and there's a lot of sort of technical arguing about whether or not it's possible to have a purely ignorant prior where, you know, we just don't we don't know anything and we have a uniform prior everything and Bay and I think

Keynes had wrote an argument saying it is possible to have this sort of ignorant prior, like a uniform prior over everything. And so that was in, I want to say general theory, but at this point I'm sort of just guessing, you know. But yeah, Keynes definitely was fully aware of Bayes and I think was a defender, but I'm now nervous that I've completely got him backwards and he was actually an implacable opponent. So who knows, you know.

Ryan
Do you have any sense for how much it affected his broader worldview?

Tom Chivers
This is where I should be up more up on family history than I am. But the, I think he, I mean, short answer. No, I don't know how much directly influence him, except that, I mean, you, he was, you know, he was a utilitarian, he was a economist and had, if you're in economics, you can't really do sort of utility calculations without Bayes, right? You need it for,

Ryan
You need priors.

Tom Chivers
Yeah, exactly. You need to have priors and you to up there when you were saying how likely something is to happen. you know, all the stuff, utility times, probability is your expected value and all this stuff. It, it, it, it need, you need Bayes as part of the calculation for that. So it will be very surprising to me if he hadn't sort of incorporated that into his thinking. I mean, it, it was really a bit later than, cause you know, the real work on, utility theory.

really fell out of sort of John von Neumann and Morgenstern. And that was, I think, in the Second World War, towards the end of Keynes's life, and probably wouldn't have become a made, you know, really... I might have got my timeline all muddled up here, but I think that's true. I don't think Keynes would have been fully on board with sort of mathematical utility stuff, because I don't think it would have fully, like, overlapped with his lifetime. But the sort of the fundamentals of Bayes would have been very familiar to him. And I hope, because I think it...

It's a really useful way of looking at the world. I hope it would have influenced him.

Ryan
Do you feel like in a sense you're carrying the torch of the scientific rider for the family?

Reply

or to participate.