Scott Aaronson: The Greatest Unsolved Problem in Math

Scott Aaronson joins us to explore quantum computing, complexity theory, Ai, superdeterminism, consciousness, and free will.

TIMESTAMPS:
- 00:00:00 Introduction
- 00:02:27 Turing universality & computational efficiency
- 00:12:35 Does prediction undermine free will?
- 00:15:16 Newcomb's paradox
- 00:23:05 Quantum information & no-cloning
- 00:33:42 Chaos & computational irreducibility
- 00:38:33 Brain duplication, Ai, & identity
- 00:46:43 Many-worlds, Copenhagen, & Bohm's interpretation
- 01:03:14 Penrose's view on quantum gravity and consciousness
- 01:14:46 Superposition explained: misconceptions of quantum computing
- 01:21:33 Wolfram's physics project critique
- 01:31:37 P vs NP explained (complexity classes demystified)
- 01:53:40 Classical vs quantum computation
- 02:03:25 The "pretty hard" problem of consciousness (critiques of IIT)

NOTE: The perspectives expressed by guests don't necessarily mirror my own. There's a versicolored arrangement of people on TOE, each harboring distinct viewpoints, as part of my endeavor to understand the perspectives that exist.

LINKS MENTIONED:

[TheoriesofEverything]

TIMESTAMPS:
- 00:00:00 Introduction
- 00:02:27 Turing universality & computational efficiency
- 00:12:35 Does prediction undermine free will?
- 00:15:16 Newcomb's paradox
- 00:23:05 Quantum information & no-cloning
- 00:33:42 Chaos & computational irreducibility
- 00:38:33 Brain duplication, Ai, & identity
- 00:46:43 Many-worlds, Copenhagen, & Bohm's interpretation
- 01:03:14 Penrose's view on quantum gravity and consciousness
- 01:14:46 Superposition explained: misconceptions of quantum computing
- 01:21:33 Wolfram's physics project critique
- 01:31:37 P vs NP explained (complexity classes demystified)
- 01:53:40 Classical vs quantum computation
- 02:03:25 The "pretty hard" problem of consciousness (critiques of IIT)

[a.hardin]

Scott is one of the great public intellectuals of our time and a great thinker that has contributed a lot to his field. What a pleasure to hear him! I highly recommend his book “Quantum computing since Democritus.”

[TheInnovator]

Scott is so well spoken, he makes a great science communicator

[Julian.u]

Scott is one of the best minds alive. He is both deep and clear at the same time. He is devoid of woo but he still does not mind touching difficult topics.

[Elon__Bust]

Okay, this annoyed me so i'll just explain it for Aaronson and others because i feel like at this point his point of view is just misinformation:

Computational Irreducibility and Chaos are similiar but not the same. This difference is explained in NKS, but I'll explain it more deeply. Computational Irreducibility is formally, the notion that all systems are equivalent to a Turing machine (under principle of computational equivalence) and therefor is subject to the halting problem : That if you were to predict the future outcome of that system...it is arbitrarily complex, and therefor you can not predict whether that system will halt, or not halt.

In more practical terms it's obvious to see why : Subsystems in the universe, can not be isolated from other systems, and so there is an arbitrarily large number of subsystems that act as inputs, for the system you are trying to produce output for. This is why the rules in Wolfram's CA experiments emulating other rules under different initial conditions is so important in the book: It proves the deep connection that systems emulate one another because of this phenomenon of reaching out in the rule space of other rules.

This is STRICTLY a stronger statement than chaos...which is only a statement about initial conditions and not knowing them to precision...where as computational irreducibility is saying much more then that : That you can not in principle predict systems because the problem of doing that is arbitrarily complicated.

That means its not just a problem of finding out what the initial condition is for your system...it's the fact that you CAN NOT find out the systems initial conditions cause the system can not be isolated: you need information of the system, and all of the subsystems in the universe, which is both the full set of information of the past AND future (the full state-space of a true Turing machine)

[Aarron-iopm]

Started by listening to the Chris Lang podcast thinking this is pure madness, then listened to Chris + CERN Scientist, and now im staying because youre awesome Curt

[aroemaliuged]

A Michu and Tyson diss in the first five minutes …
Subcribed

[Trader-]

Scott Aaronson is a professor of theoretical computer science, but his presentation of quantum mechanics is better than that of any physicist I have ever heard. He is literally the only great mind out there that does not talk jargon, gets to the very heart of the matter, and presents his ideas in the most crystal clear way I have ever heard.

[BoRisMc]

this man is truly scary smart. Gonna need a good few passes to fully go thru all the material Scott is casually splurting out. Really nice drop!

[happywednesday]

@14:00 I actually think we do know a perfect brain prediction machine isn't possible, because similar to the halting problem and the variants, you can pose that you could have an interaction with the machine itself, and just as the halting problem cannot know itself, the perfect brain prediction machine cannot predict itself and therefore cannot perfectly predict a brain interacting with it

[ariadne]

thank you Scott Aaronson for correctly and succinctly definining superposition! As my Quantum Physics I professor said "it (electron, photon, etc.) is not either, nor neither, nor both, but in a superposition, which we don't have the language to describe."

[williamjmccartan]

Reminded me of Dr. McCoy never trusting the teleporter 41:27, thank you both for sharing your time and work Scott and Curt, Scott's enthusiasm is great to see, love his honesty regarding looking at others theories, peace

[darkbb]

❤ I always enjoy hearing Scott Aaronson speak

[almari]

One of the few honest people in modern science.

[luken]

Curt, again and again people misunderstand what computational irreducibility is, and I am extremely disappointed in Scott making the exact same mistakes. Yes it is in fact a new concept. And no, it is not just chaos.

In chaotic systems, you can always "look ahead" and predict the future state at a time t, provided you have extremely accurate information about the initial conditions.

In computationally irreducible systems, you can do no "looking ahead". It's impossible regardless of the accuracy of your initial knowledge.

Chaotic -> tiny change big consequence
Irreducible -> no shortcut calcuation

[ty]

Excellent Guest…. You ask very good questions …! Excellent interview!

[vfwh]

100% agree with his view of free will. I created exactly the same argument about free will in something like 2012, and I'm glad that Curt asked about the sealed envelope, because if freewill being an illusion depends on the fact that I can't look in the envelope, then your argument against free will is tautological by design.

In fact the argument for free will is simultaneously an argument against the core and universal determinism of the universe.

[vootman]

Nice Curt...great upload. I think it has to be P doesn't equal NP. Our universe doesn't work the other way around. I think the problem is somehow deeply tied to entropy. Like literally checking the steps of a problem never takes as much work/time/energy as coming up with a solution in just about every aspect of life....things would be easier if they did!

[IlEagle.G]

One of the best conversations you’ve had!

[dimitrispapadimitriou]

18:00 The most plausible answer to Newcomb 's paradox is that it's a kind of reductio ad absurdum for strict determinism and full predictability.