What is Trust? w/ Karl Friston

I am joined by neuroscientist Karl Friston to discuss, from first principles, what trust is. The trust we put in others, and in our institutions, is a sociological feature emergent from biology. It is here that predictions first appear. This conversation inspired an article we co-wrote: ⁠https://schweizermonat.ch/trust-is-a-law-of-nature/ 0:00 What is trust, in active inference terms? 1:07 Trust in your own future: planning as inference 2:16 Confidence, precision, and the shape of a belief 3:45 Two ways to minimize surprise: seeking info vs. avoiding surprise 5:48 Epistemic trust: choosing where to get your news 6:24 Scale invariance: from eye movements to which institutions you trust 8:20 Trust as a meta-judgment of your past predictions 9:11 The statistician's view: estimating risk, not just reward 10:56 When second-order inference fails: hallucinations and neglect 12:50 Delusional institutions: the same mechanics in markets and geopolitics 13:45 Precision as attention: what you trust is what you attend to 14:56 Rediscovering natural selection from a different angle 17:20 Bayesian model selection is nature's way of selecting phenotypes 19:58 Why the free energy principle complies with itself 21:14 Donald Hoffman, mortal computation, and neuronal chauvinism 22:31 We are the substrate: predicting ourselves into existence 24:07 Everything is a hypothesis, including time and the self 25:10 Do we see truth? What can be ignored isn't true 26:13 There is no truth, only the best inference for me 27:33 Synchronization of chaos: coupling your brain to the world 28:45 A joint free energy solution: becoming mutually predictable 29:38 Aligned interests: why you trust the table, your liver, and me 31:52 Negative valence: trusting that a lion wants to eat you 32:47 Shorting the BBC: the cost of distrust 33:33 Fake news, epistemic trust, and the pillars of autocracy 34:15 Sharing a generative model: how we become predictable to each other 35:47 Defining yourself by the states you refuse to occupy 38:04 Is this source generated by someone like me? 38:36 Why echo chambers are Bayes-optimal until communities open 39:54 Too many choices: why it's awful for young people now 40:58 Friston: no phone, no social media, no awful uncertainty 41:15 Australia's under-16 social media ban, seen mathematically 42:06 Information like capital: who should handle how much 42:48 The pathology: social media attracts more capital when you lose 43:28 Markets are grown up, social media is pre-adolescent 44:11 Science as selection: theories as surviving hypotheses 44:35 Is religion "true"? The fitness of a simple generative model 46:03 Narrowing choice: why fewer options relieves angst 47:05 "You are your own existence proof" 48:07 Collapsing is and ought: first survive, then philosophize 49:08 "I am, therefore I think" 50:23 The relational aspect: no standalone measure of goodness 53:30 Hardware and software differ only in speed of change 54:33 The renormalization group: same rules at every scale 55:07 Separation of temporal scales: growing slow, learning fast 56:15 What we inherit is learnability, not a finished world model 57:40 Why foolishness was selected for: the reckless years 59:21 "Did evolution fool us?" The value of unsophisticated agents 1:00:24 Variance creates arbitrage: Pinker and common knowledge 1:01:25 Closing thoughts