Is Bayesian thinking a sham?

There's a common criticism that people who claim to practice "Bayesian thinking" aren't actually doing anything special -- they're just using a fancy term to make their opinions seem more objective than they really are. In this video I explain why that criticism is misguided.

LINKS:

[jetwim]

Julia has the ability to speak cogently for a long time without a single edit and it is high sorcery

[Liphted]

I can't wait to catch her new video 5 years from now!

[dande]

Evidence points toward's Julia's uploads being beneficial and greatly apprecaited. I hope she continues to act according to the beliefs upheld by such evidence.

[allengreg]

Words have the power to both wound and heal. When used with kindness and empathy, dialogue becomes a bridge that connects hearts and minds, turning pain into understanding and loneliness into connection.

[Tango_Alpha_Charlie]

Hi from Nevada. I'm enjoying the process of learning from your videos. This particular one gives me new appreciation for the value of acceptance of uncertainty. It also makes me better appreciate the approach used by British Intelligence. My exposure to their statements began as I was reading about their take on the russian invasion of Ukraine, and the Ukrainians fighting back. British Intelligence goes out of its way to be open as to the level of uncertainty underlying a particular statement, including publishing a list of definitions of the various words they use to convey a particular level of uncertainty. I like that, by implication, they see value in publishing information even if they are not 100% sure about it. Rather, they include how certain they are. If modern-day "me" could visit 18-year old "me" then good advice to give would be that an idea doesn't go from useless to very valuable as its probability goes from less than 100% to 100%. It's more of a gradual thing.

[danfg]

I don’t know... the cognitive dissonance of supposing I’m wrong just feels like inviting a lot of trouble, compared to the warm feeling of being validated by anecdotal evidence.

[bocckoka]

I'm not a Bayesian, I'm just lazy, an I just dismiss statistical claims based on individual cases instantly. This way I have time to build little ships inside bottles as a hobby.

[thomaskrenn]

i love humans who want to think critically, openly, apply it and communicate it to us. I feel comfortable here.

[KarlLew]

It is really really hard to put the world at the Bayesian distance required for rational discourse. Thanks again for a great video!

[jd]

Julia, PLEASE DO A VIDEO ON AXIOMS! you are so good at explaining this topic. Probabilistic reasoning and inductive/deductive reasoning would be a great tone too. Thank you so much for sharing your knowledge with us!

[oomenacka]

The words "I don't know" are very powerful. We all need to learn them.

[le_scienceall]

Thanks for clarifying this! This is helpful!
About the word "Bayesianism", in my book "The Equation of Knowledge", I distinguish "pure Bayesianism" and "pragmatic Bayesianism".
A pure Bayesian rigorously applies Bayes rule upon any evidence. While impractical, I find it to be a nice thought experiment to convince myself that, yes, in a world where computation comes for free, this is the way to go.
In practice, we must then resort to being "pragmatic Bayesians". This means that we strive to find good heuristics to approximate Bayes rule. Interestingly, this applies to machines to, and there's a lot of surface tensions between AI research approximating Bayes rule and what Rationalists try to do 😄

[jimnesstarlyngdohnonglait]

Think more the hardest way as how it can be noticed that at the end of the latest Sentence in the previous comment, two separate quotation marks appear: a single quotation mark for closing the title and a double quotation mark for closing the direct quotation.
Vs
Think more the Hardest way as how Lewis Caroll said, "Always speak the truth, think before you speak, and write down afterwards."

[jimnesstarlyngdohnonglait]

Think more the honesty way...
Vs
Think more the hardest way always

[larshelmin]

"All models are wrong, some are useful" - George E P Box. A Bayesian model of the world is useful I think.

[budjohnson]

Wow, how enlightening for me, first time on this channel I think. Almost 40 years ago I was early in college taking paleontology as a part of my degree in Geology. I brought a fossil to class I found on the ranch to learn its origin. The prof asked the age of it. I said I don't know because it was found on an unconformity (either 65 or 225 million years ago, Cretaceous or Permian) but probably the younger because rocks roll downhill, turns out I was right and prof said I would be a good geologist because I started with "I dont know" and discussed probability of either. I was surprised by this because by age 19 that statistical way of thinking was completely ingrained, I didn't know another way. I recall similar stories from a much much earlier age. So assuming this is some sort of Bayesian way of thinking, the interesting question is, what is the evolutionary origin of this type of thinking. I suspect it is an inherited trait that is either reinforced or quashed at an early age. And maybe they are selectively chosen by our parents and peers at an early age which explains the difference in our common reactions to weather or politics? What a great video, thanks!

[Sayerik]

"Why are you glaring at me like that?"

Julia said I could.

[jimnesstarlyngdohnonglait]

Think more to never look down...
Vs
Think more to always look up

[jimnesstarlyngdohnonglait]

Think more when I learnt how to hate what I loved curious about just like my past old days
Vs
Think more when I learnt how to be satisfied in the warmth of your body

[Richard_Jones]

It should have been really obvious but the idea of saying to oneself "ok, what if I'm wrong" and treating that idea seriously is actually mindblowing.