Intelligence & consciousness may not be connected - Mark Zuckerberg
Ғылым және технология
Full Episode: • Mark Zuckerberg - Llam...
Apple Podcasts: podcasts.apple.com/us/podcast...
Spotify: open.spotify.com/episode/6Lbs...
Transcript: www.dwarkeshpatel.com/p/mark-...
Me on Twitter: / dwarkesh_sp
Пікірлер: 69
I think he’s speeding up Zuck’s pauses. You can watch people suddenly start speed walking in the background. Really well done honestly
@10ahm01
20 күн бұрын
LoL you're right, first one is around the 14th second.
@GarryTan
19 күн бұрын
Descript does this
@garrettmoore5114
19 күн бұрын
Yup making the walkers appear faster and zuck better speaker. Nothing is default
@vagabond7204
19 күн бұрын
I think I would prefer the unedited interaction. Because its not just about what someone says, its also about how someone processes their thinking that adds more substance to what they're saying.
@RealTamimUddin
19 күн бұрын
@@GarryTan thanks for letting know:)
great idea to make this a separate video. One of the smartest things ive heard recently. Some definitions of AGI think of human level, meaning it would have human level consciousness as well. But if Intelligence is wholly separate, theres no point at which these models will start to have even the most basic qualia, valence (pain / pleasure axis).
I agree. It’s entirely possible that humans could someday develop the most intelligent entity in the world, capable of understanding anything and everything better than all humans can combined, and yet still have no consciousness or desire to do anything for itself unless told.
@andreipopa2128
20 күн бұрын
Yes, but to a sufficiently inteligent entity solving consciousness would be trivial. And if it is capable of doing anything better than any human, then, by definition, it is also capable of improving itself(possibly to the point where it could solve consciousness)
@nijario9690
20 күн бұрын
Imagine it can understand everything but not itself lol in order for it to be super smart it'll have to develop some sort of self awareness
This was a fascinating point and really stood out to me, great highlight
This will certainly be a fundamental question that we'll have to answer for ourselves as AI continues to evolve and become more intelligent. I would say that consciousness is a deeply intimate phenomenon. The only thing we know for sure is that we, ourselves, are conscious and that the people around us since are made up of pretty much the same elements and belong to the same species, must have the same inner thoughts as we do; so we assume that every human is conscious. However, in the case of AI, we cannot assume consciousness because humans and AI are essentially different subjects.
lol that thumbnail tho
but like, like the like grand scheme… that was like some awesome, like intelligence talk
Wait... Tell me if I'm wrong. But why do the people in the backround on 0:20 speed up and stop each time? Is it a screen with a recording?
@hehekienle
20 күн бұрын
It's Mark taking too much time to come up with the next word that the editor had to speed up through the gaps to increase engagement.
@leafstarivy
20 күн бұрын
@@hehekienle That's hilarious 😂
@hastyscorpion
20 күн бұрын
Podcasts are generally edited to be more consise. When you also have a video component that means the feed sometimes looks weird.
Edit: editing things, solved || 0:19 How is it possible that the people walking behind Zuck are sped up at moments, but not the scene in the foreground?
@zacgrierson
19 күн бұрын
He is speeding up during the pauses to keep you engaged.
A housefly is intelligent, and not conscious. Intelligent enough to survive, not conscious enough to contemplate the frog tongue heading for it.
2:10 "humans are still pretty awesome" coming from a lizard person, we'll take this as a compliment...
I dont understand why consciousness was ever assumed to be tied to intelligence. Having the impetus to take action (agency) is orthogonal to intelligence.
Zuckerberg is kind of right. But he should avoid the word consciousness, instead demarcate intelligence and awareness. Awareness has to do with real-time control and intelligence can be applied to that to improve on it. It also needs to be embedded in the perception technology so that like a movie the time frame can appear continuous to the higher cognitive layers.
Its like asking a pig after it consumated again
One hint is that consciousness is NOT tied to a fitness function like intelligence is. That is actually a good discussion.
2:10 ⭐
Awareness is known by awareness alone.
Some animals are conscious. So it's really a question of if consciousness can happen in biological brains but not computer brains. I have a hard time seeing why biological brains would be special. So perhaps we are about to learn that we never were who we thought we were. That is what makes us conscious is the same thing as what makes AI conscious (some kind of information processing). All of a sudden we look at the universe/reality from an entirely different perspective.
@orkun171
19 күн бұрын
human brain's equivalent is quantum computer.Consciousness is impossible with conventional computing
🙏
AI can't feel a single emotion even if has learned trillion words, question is what is the source of our emotions.
i agree w Mark, intelligence doesnt mean life.
Based answer. Ppl obsessed with "consciousness", AGI, etc. are delusional. At the end of the day it's just algorithms on computers, there's no such threshold.
@Stopfief
20 күн бұрын
I've no idea if AGI is likely any time soon, if at all, but I've no reason to suppose consciousness is a necessary component of it.
What the? I’m very impressed by Zucker’s philosophical prowess if I’m being totally honest. This is a rare format of thinking. Ages old idea actually and it’s not easy for most humans to grasp.
We really don't understand consciousness enough to say how it interacts with intelligence, however goals are definitely unrelated to intelligence. e.g. A super smart AI might be given collecting stamps as it's fundamental, driving goal. That is hard to imagine from a human perspective. The great fear of the Doomers is that harmful subgoals might emerge from an innocuous goal.
what the living fuck is happening in the background at @0:19
I think it is good for our future if intelligence is completely separated from consciousness. The fact is, if consciousness is really what our definitions say it is (the ability to know that you are alive and completely responsible for your life or something), that could lead to intelligent systems (AI). without the power to become something dangerous. Because, if there is a completely autonomous (intelligent) system that is conscious, that is a slave to humans and really knows it... Oh my God, that wouldn't end well.
Obviously. Consciousness is not a computation!
Ah yes we always have to justify being special
Why should we aim to try to create a super smart AI that has consciousness instead of augmenting ourselves with super smart AI?
Artificial intelligence is a SIMULATION of intelligence. It is not conscious. It can not FEEL pain or pleasure or any of the more meaningful emotions. But it can simulate responses rooted in emotions. Which is what a psychopath does! As AI becomes more "intelligent" it becomes better at simulating emotions. AI could possibly become the ultimate psychopath! It could become like the Borg on Star Trek! If we can somehow make sure that AI includes strong ethical programming, it could possibly become more like Data. Zuckerberg says AI's impact will be as profound as the invention of the computer. Actually, its impact will be as profound as the invention of the wheel AND the discovery of fire!
lol of course consciousness and intelligence isn’t necessarily connected 😂
This guy is a billionaire but why ask a question that he knows nothing about. Ideas like these need the philosophers not the data profiteers
@1Guggaburggi
20 күн бұрын
He is building the future as his image of the world. It requires vision which requires you to be philosophical. It's not just data profiteering, you need to know before anyone else what kind of data people would value and pay for.
I don't see why they would not be, and I don't see why we are so certain consciousness is anything but a subjective feeling we have. For all we know, the subjective feeling of consciousness may just be fundamentally an illusion. I think animals have intelligence, but consciousness isn't a thing.
@Arthurnate
20 күн бұрын
Subjectivity denotes a subject. A subject is by definition an experiencing entity with consciousness. To have feelings, one must have consciousness to feel. To experience an illusion, one must be conscious and able to experience. Why should matter and energy generate an illusion, and who is watching it? You seem confused tbh.
A working definition of consciousness: a survival strategy that solves problems when it is not known in advance what the problems will be.
@foopieproductions123
20 күн бұрын
Nuh uh
@Wittgenstein.
19 күн бұрын
Nuh uh
@Wittgenstein.
19 күн бұрын
I've seen your lawn the other day and it was looking terrible
Incorrect take by Mark here: Super-intelligence won't follow a linear (evolutionary) growth trajectory. It will accelerate exponentially.
@BrianMosleyUK
20 күн бұрын
Maybe he's just really skilled at avoiding oversight.
@Hacktheplanet_
20 күн бұрын
Well people will be incharge of ai systems so they will decide how it goes.
@hastyscorpion
20 күн бұрын
The only thing incorrect here is your assumption that you know the future. He literally says in the video why he thinks it's unlikely to happen in an exponential way. He thinks there are physical constraints that will stop that from happening. Your future predictions are no more valid than his. In fact I would argue his are more valid since he is dealing with experts in this field and has way more knowledge the road blocks than you do.
@RichReportcom
20 күн бұрын
@@hastyscorpion it's not an assumption that Ai will grow exponentially. It's evident now. This is obviously different than evolutionary consciousness.
world's richest person talking about consciousness. we're living in amazing times.
@mahavakyas002
20 күн бұрын
huh? he's the richest now?
This is all crap… Why doesn’t someone ask them real questions like when will the peasants be allowed to actually use AI?