Theorem of Human Unpredictability - Seth Lloyd

Source - serious-science.org/theorem-of...
What are self-contradictory statements? Is there a computer that has the capacity for self-reference? Professor of Mechanical Engineering at MIT Seth Lloyd explains why no decider is able to know what decision they will make.

[@ridealone7933]

I much prefer Seth Lloyd over people like Sam Harris. Seth is brilliant, but he is smart enough to know that he really doesn’t know. The most important attribute that an intellectual needs to cultivate is being humble. Second most important attribute is learning from the past. You can see that Seth has worked on both attributes.

[@golagaz]

They are not comparable. Sam Harris is social scientist compare to Seth .

[@abhishekshah11]

Sam Harris is a girl. I don't like listening to him.

[@KipIngram]

I agree with you about Harris - there are several of the personalities out there right now who project a smug "I'm so sure of my rightness that it makes me superior to everyone else." Sean Carroll is another example.

[@malayangrago5628]

Truly a genius. 21st century Einstein. Got to watch more videos on Seth .

[@user-cu9ww9tj4i]

휴리스틱 사고를 하고 잘하는 사람도 있고 못하는 사람도 있다고 대략적인 편견없이 그냥 느끼면 그나마 편견이 적은 것같아요.

[@solearv]

Just brilliant

[@luckyyuri]

+Antón Covariante free will is such a big, complex, encompassing term that it's impossible to be approached by science! i like lloyd but here he's behaving like naive folk. if someone spills a drink on you, instantly(!) you get angry... where is "YOUR" decision there? you might next time decide to repress anger to show you have free will, but you still burn your neurones inside with anger and high pressure :)) that is not free! anger is highly toxic for your organism but however most people are in constant conflict with the most trivial things; see how you manage that piece of information. ever heard of "pushing someones buttons"? that saying has come up because we are in some degree mechanical in behaviour, like fans! also did you decide who you are? i bet if a fairy asks you if you want to change something about yourself you would jump in microsecond wanting to have better control of emotions, higher logical capabilities, higher memory recall, higher working memory space, funnier, capable of higher outputs of happiness, love, orgasm, etc etc... once transformed and having better understanding of things and having other feelings toward things, you might want other things all together. and there is also the problem of "who is having the free will?". who is you? people don't have an identity, and authentic inner core; all they are is a set of "ready-made" stuff that are copy/pasted for their overwhelming majority by the unconscious. my neighbours are gypsies and their little children copy the behaviour of the adults like f... xerox machines :)) if you would take one and raised it in the saxe-coburg-gothe family he would be appalled if he would see his brethren on the streets! where's the free will in choosing the most important thing... WHO YOU ARE?!! and so on and so forth, i could go on till next week. "people are puppets in their behaviour, but they become poets when they describe it" this is a piece of geniality from the old days, osho if i remember right

[@endar117]

+anywherein12seconds I didn't read that^

[@ridealone7933]

I much prefer Seth Lloyd over people like Sam Harris. Seth is brilliant, but he is smart enough to know that he really doesn’t know. The most important attribute that an intellectual needs to cultivate is being humble. Second most important attribute is learning from the past. You can see that Seth has worked on both attributes.

[@KipIngram]

0:30 - THAT'S NOT AN ARGUMENT. When people thought the laws of physics were completely deterministic, THEY WERE WRONG. Being wrong about something doesn't prove anything. In fact, I have long thought that as soon as that idea started to set in, someone should have immediately stood up and said "This can't be right, BECAUSE it doesn't allow free will." Long before quantum theory came along and threw determinism out, it obviously had to be wrong. I regard the self-evidence of free will, and the self-evidence of our own self-awareness, and "more primary" than experimental scientific evidence. All of that evidence comes to you via your sense organs and has been subjected to a great deal of neural processing. These sensations of free will and self-awareness are more direct than that; they're as direct as anything can possibly be.

[@eriklagergren]

First, there is a difference between assesment and computabillity. The haltingproblem concerns the later. Determening an outcome with a certain magin is not restricted by the haltingproblem as the depth of the calculation often say something about the probability. Second, there is a difference between option and probabillity that Lucretius and many others missed. The probability can in some cases be virtualy one to choose an option and stil most people would consider it to be a free choice as long as it is deliberate and fair.

[@KipIngram]

5:15 - You used your free will to choose to flip the coin. So far I'm not getting much out of this - I agree with the guy who said "Our will is free and that's the end of it." The problem here is that it's impossible (so far as I know, and I've thought about it quite a lot) to *prove* that an action was an act of free will. No matter what you choose, someone can say "That was just what was always going to happen anyway." You can't "re-wind the universe" and do it the other way the second time. BUT - failure to prove a claim does not constitute a proof that the claim is false. Science should have NO POSITION on free will, because you can't prove it exists and you can't prove it doesn't exist. It is not in the realm of science at all. The same on the existence of self-awareness. These are things that we all directly feel, but there is no way for us to "project that feeling" into someone else's mind. All we can do is use spoken and written words to write down a claim about what we're feeling. But that's not evidence of anything.

[@ibperson7765]

Isnt it part of the definition of decide to say if you know what youll decide then youve already decided

[@diego898]

fantastic video! Is there a reference for the theorem hes proved for any physical system?

[@luckyyuri]

+Diego Mesa consciousness may be (who knows?) some metaphysical substance, but the mind inhabited by that consciousness definitely acts by classical rules of determinacy. there's a definite correlation between brain states and mind states, seen through endless medical experiments. so, at least, mind behaves in synchronicity with the underlying physical events of the physical brain... and the brain is a macro system that evolved in a macro environment, with evolutionary pressures to react to classical interactions in that environment, as faithfully as possible! < if i push a ten meter rod by one end, i don't need to watch every quantum aspect of it's inner structure to know that the other end would move in accordance - classical logic! > brain is a machine that does everything in it's possibility to eliminate randomness from it's behaviour, because in the course of evolution, billions of organisms (from the first nervous system) in their countless interactions with a classical environment, the brains that didn't interpret things in the deterministic logic were not fit to pass their genes! all people living today are in a select lineage of highly faithful brains to their environment! for example: i'm the clone of my father, it's incredible how we look alike, cough alike, we sneeze alike (always twice), sleep alike, we have similar traits of character... but we think so fucking different! the former, i think are traits that have to do in considerable amount to genes, and the latter has to do with the environment! we are what we see! my brain and his, were programmed by different environments, and as i sed above, brains are machines that grow faithful to their environments! i'm just what the environment did to me, and what's done stays securely done in a way that only another classical interaction could possibly change my behaviour (improbable and very hard if i'm over 20). free will is clearly an illusion because there's no free will in choosing who you are and how you act and react in the first place. seth lloyd is only discussing about the narrowest frame of the concept, at the very moment of decision making - where it doesn't even matter... because of course here the decisions are arising free, nobody said there's a puppeteer or a magician playing with our neuronal paths. he's leading people to wrong impressions; and boy did we jump into conclusions when they flatter our egos

[@yazicib1]

Randomness (if it actually exists) would not imply free will. Proof of intrinsic randomness in quantum level is also impossible to prove. Sentence "I make a decision" has a bigger issue than free will actually. It is the word "I" and exact definition of it in the physical sense. Free will is a metaphysical concept and cannot be the subject of science as it is not provable or disprovable. So is randomness. Unpredictability, on the other hand, is obviously a fact due to our inability to know the current state to predict a future state (i.e. Heisenberg's Principle)

[@luckyyuri]

+yazicib1 brilliant! thanks for making it so clear

[@user-cu9ww9tj4i]

상식적으로 생각하면 우리는 우리가 태어나고 우리가 태어나기전 모든 정보를 바탕으로 모든 것으로부터 합리적으로 생각하는 것은 느낌상 안될거같기도 합니다.

[@sparhopper]

11:58 = the most important statement made here.

[@Novak2611]

Let's assume that neuroscientists have a device that predicts the choice of Lloyd 1 second earlier. So he can also use the device to predicts 1 second earlier what the neuroscientists will think about his choice, that is Lloyd will know what will be his choice 2 seconds earlier and so on, Lloyd would know the choice he will make an infinite time earlier which is a contradiction, unless time has no meaning in reality, which makes it then a useless tool to understand what the free will is.

[@LaureanoLuna]

Interesting. Lloyd's wife uses the Device to predict Lloyd's behavior while at the same time Lloyd uses (another token of ) the Device to predict the behavior of her wife, hence also his own behavior; but once he foreknows it, it can surely change it, making the Device fail.

[@davidwilkie9551]

The definition says, in short, "the power to act without constraint", so the binary examples given don't fit the circumstances, where an individual can chose to ignore the offered choice altogether. If you have a sweet in each hand and say to a child, "you can have whatever you want, it's your choice", what's to stop them taking both hands? The computer can also categorize an additional functional computation to be processed according to an alternative choice to completion of the set one. It's the same, everyone does it all the time, and it's nothing to do with intelligence, only free will.

[@Novak2611]

We can also notice that in a consistent system, if we consider the sentence: S="This sentence is not provable". Then for a human mind, "S" is true. But a computer or another algorithmic process cannot decide whether it's true or not. This shows a fundamental difference between minds and machines. We can also observe that for example in the fact that mathematicians can play with infinities (for example "The Axiom of choice") to derive results, which is not possible for a computer.

[@LaureanoLuna]

For any system S, a system S* exists that proves "this sentence is not provable in S".

[@donkeychan491]

If the iphone implemented a hard-coded rule , and knew that to be the case, then surely it would be able to predict its own decision? I'm unclear what the distinction is between a "decider" and "non-decider". Are "deciders" defined to be (sufficiently complex?) computational entities that can't predict their own future decisions?

[@KipIngram]

10:20 - That would only apply to humans if our minds are in fact just programs running on a computer. That's not the case, though, so the argument is inapplicable. Your theory is probably still right though - I happen to NOT believe that our minds are merely a mechanism running on a physical infrastructure. Computers will never be self aware, either - for the same reason. Whatever that "more" is, computers don't have it. They ARE merely processes unfolding on a physical platform.

[@KipIngram]

2:08 - Before you DO THE ACTION. That is not the time you made your choice. They have no IDEA when you made your choice, so these experiments are utterly devoid of significance.

[@ProtectTheMessage]

When I started to watch this I was trying to decide if I would watch it all. Before I could decide I found myself watching it to the end. Having not made the decision, was it my decision to not decide or did I decide to watch it all by some default 'non-decision'? If I have no free will, then what would be the functional reason(s) my brain would trigger my initial 'watch it all or not watch it all' question?

[@rovrola]

Resource management. If an organism evolves with limited access to resources and significant competition, modelling future outcomes and decision-making would seem to be useful adaptations. Particularly tricky situations might inhibit other physiological processes or habitual neural signalling externally manifested as the individual pausing to deliberate. This deliberation in itself is an expensive process, but it might eventually become a state worth identifying, eliciting, defending, etc. by the community of agents.. and "free will" is born. Perhaps a long series of accidents programmed free will in us? :) I see free will as part of a social mechanism involving responsibility, justification and identity construction, emerging as a resource management function among individuals as well as within.

[@hoppechr]

Very nice talk and very sympathetic researcher. However, unpredictability is irrelevant for free will - our will could be free even if we knew our future decisions, why not? We would just have made the decision much earlier. In fact, in many important decisions an early though not yet definite "intuition" of what we are probably going to do can be perceived and often also remembered. Unpredictability, however, provides a very good basis for the illusion of free will and for assigning free will to others. I see no reason to introduce any new source of movements besides the known four "interactions" or forces (i.e. weak, strong nuclear force, electromagnetic force, gravity/mass). This leaves neither need nor space for truly mental causes.

[@charleskoenjgsart4712]

We do have Free will , which is definted by frontal cerebral-cortex , but our brain also have a system that records information , so when we face the same or similar situation ,we do not have to solve it from scratch , so that we can have time dealing with further problems or improve ourselves or other things . And people who have lived in bad habits ,is their own choices that have been made by themselves that makes them living in this abyss , the hell-like psychical state ,and by the way, there is no heaven or hell , it is just an analogy of mental state...for positive and negative.... We ,human beings do have free will , and which is why we are truly responsible for our own choices and the related behaviors and actions, instead of god or gods or any other fabulated characters...which are only imaginations and ways to excuse one's own irresponsible doings...

[@rovrola]

Is a UTM simulating another UTM computationally the same problem as a UTM simulating itself?

[@reinerwilhelms-tricarico344]

Good question: I might be the same, but depends on how the two machines are interacting in real time. It depends on whether or not the two systems are completely decoupled or not: Machine B is trying to emulate machine A: First, what is going on in B is quite different from what is going on in A: B carries out the same operations that A does plus a lot more... When B starts the computation it would have to know the state of A in the beginning of the entire computational process that leads to A's final decision, and it would have to know exactly the input that A receives. If A and B are separated after a certain point in time, and none of the computation of A (or B emulating it) depends on any further to be observed input, than it should be possible that B (if it's faster) can predict A's decision. But if both the processing in A and B's emulation of A depends on further input (meaning A is not closed off) and B receives the same input as A at the same time, then one would think that in principle B is able to predict exactly what A will decide - unless B finds out that A's decision processes are completely independent of future input. Otherwise, B could not finish that computation until A is done - Because, simply, it has to wait for even the last piece of input that A receives. B can't anticipate A's next step if it depends on that input if B doesn't know that input. Otherwise, since B doesn't know the future input to A, it can't predict what it will compute: It's own emulation of A depends on that missing information. So in other words, unless B knows both the initial state of A and all inputs that A receives up until the moment it makes the decision, B can not anticipate what A will decide. However: It can fully anticipate what A decides in the end - but only after the fact. It can make excellent predictions about the past but not about the future. When A then comes up with its decision, B might then say: I knew you would do that - but it would be lying - it had figured it out only after A did it. I just realize how I slowly begin to slip into anthropomorphic thinking ... so I better stop. OK, just my two cents worth, when I try to think about computation as actual physical processes. I may be wrong.

[@reinerwilhelms-tricarico344]

Just realise it's even more complicated if any of B's actions while trying to emulate A have an effect on the environment of A, and thus possibly contribute to changes of the input that A receives. In that case B would be confronted with the problem to also have to simulate the environment, namely A's environment. And that would include B itself. In other words if the two systems are in the same universe and interact in any way, neither A nor B can make a prediction with certainty.

[@briansmith7458]

Anyone know where to get good acid?

[@amanda3172]

kek phobia is real

[@noahway13]

You can legally grow magic mushrooms. It is the next best thing. Google it.

[@billbodge3879]

Free will in the sense that we have a choice among a set of probabilities.

[@LaureanoLuna]

He seems to be asssuming the controversial claim (controversial to the point of unlikeliness) that human agents are equivalent to formal systems and Turing machines, so that Gödel's and Turing's theorems apply to them.

[@alexvandenbroek5587]

Isn't it simply the concept of a physical system being able to compute or process information. That doesn't seem very unlikely unless you assume there's some special magical floating free will substance.

[@LaureanoLuna]

@@alexvandenbroek5587 It is indeed very unlikely: algorithms (due to their very nature) are completely objectivated for the human mind but the human mind cannot be so for itself, otherwise it could go beyond itself. Take a look: www.argumenta.org/wp-content/uploads/2016/12/Argumenta-21-Laureano-Luna-Russellian-Diagonal-Arguments-in-Metaphysics-2.pdf

[@aminkanji8501]

After listening to this I feel like drinking 😌

[@judgeomega]

Decision and the halting problem are very separate concepts in my mind. I think it is logically fallacious to claim 'proof' of determinism/freewill this way. Its almost immoral to inject such a falsehood into the profoundly important subject, at least without admitting the limitations of the analogy. Without a mathematical definition of; 'decision', 'free will', and 'determinism', we will continually struggle with such ambiguous vagueries. Human language is inherently flawed. And using flawed tools will always result in crude results until will make better tools; a more scientific language free of contradiction, self reference, and fallacies.