Re-reading Nassim Nicholas Taleb and I can't not notice how his majority of ideas are commentary on Richard Dawkins ideas of evolution. I call this Tracer bullets theory - nature/entrepreneurs can't predict future but fire many tracers (individuals/startups) to find targets.

FED charman Progress so far

All systems are causal mechanisms, and all causal mechanisms are systems.

If everyone adopts Nassim Nicholas Taleb philosophy of optionality, being outside of the ring, waiting for rare event to collect payoff - GDP would collapse immediately. It's not a coincidence that term "Levantism" had negative connotation back in the day in Serbia.

With the rise of AI humanity will shift focus from learning to discovering. “Neural networks just wants to learn” They are exceptionally good at learning existing knowledge, but they’re not good at discovering new knowledge.

You can now wishlist FED Chairman game on steam store.steampowered.com/app/3691190/FE…

What all people miss is that generalization can not create new knowledge. New knowledge is created by doing and based on that establishing causal relationships(Sys1). You then take this small particular part of knowledge and extend it via generalization ie. make it universal

Neural networks are not System 1 because they lack very important feature - the notion of causality.

Faculty of judgment is mediator between System 1 and System 2.

What is Understanding? It's the apprehension of direct causal relationships. Transition from effect to cause and from cause to effect, nothing more.

How should one decide which sort of representational system to use in connection with a given problem? What is that search-reducing heuristics? It's the principle of sufficient reason which reduces all problems into four categories. Causal Math Abstract Knowledge Human Actions

"The critical attitude is characterised by the fact that we try not to verify our theories but rather to falsify them. Verifications are cheap: they are easy to come by if one is looking for them." - Karl Popper

LLMs are not fraud. But the obsession with intelligence, whether artificial or not, on the other hand, is. LLMs are wonderful autocomplete tool.

"Of all the discoveries that have been made concerning the inward structure of the human body, never one was made by conjecture. " -Thomas Reid

Enemies of human progress: 1) Conjecturists (David Deutsch , Karl Popper Quote, Conjecture Institute ) 2) (Artificial) Intelligence community Common thing for all is that they believe that new knowledge is result of intelligent people conjecturing out of thin air.

Biggest issue with YC Combinator is that they don't fund domain experts. They fund skinny kids with no experience in the domain. That could work in past era, where programming was equalized with domain expertise as it was biggest bottleneck.