Science & Technology
Will Superintelligent AI End the World? | Eliezer Yudkowsky | TED
Decision theorist Eliezer Yudkowsky has a simple message: superintelligent AI could probably kill us all. So the question becomes: Is it possible to build powerful artificial minds that are obedient, even benevolent? In a fiery talk, Yudkowsky explores why we need to act immediately to ensure smarter-than-human AI systems don’t lead to our extinction. If…
Science & Technology
What TED Will Look Like in 40 Years — According to Sora, OpenAI’s Unreleased Text-to-Video Model
Sora, an unreleased AI model from OpenAI, generates realistic and imaginative video from text prompts. Acclaimed filmmaker Paul Trillo was granted early access just before the TED2024 conference. This year marks TED’s 40th anniversary — but instead of looking back, we asked Sora to show us what the next 40 years could look like. (Footage…
CNET
3 Hidden AirPods Features
Let’s try a few of the coolest tricks you can do with your AirPods 2, without a skateboard. #airpods #headphones #audio #apple #earbuds #music #howto
Science & Technology
Rabbit’s R1 and Humane’s Ai are rivals for a market that might not (yet) exist | TechCrunch Minute
Is hardware fun again? Two new AI gadgets — Rabbit’s R1 and Humane’s Ai Pin — are finally heading out into the world in an attempt to redefine how we interact with the digital world. But will Rabbit’s R1 and Humane’s Ai Pin truly replace our smartphones anytime soon? For Humane’s Ai Pin in particular,…
What TED Will Look Like in 40 Years — According to Sora, OpenAI’s Unreleased Text-to-Video Model
Investors Pleased with Cloud Results: Frisch
Bloomberg Technology 04/26/2024
-
Science & Technology4 years agoNitya Subramanian: Products and Protocol
-
Wired5 years agoHow This Guy Became a World Champion Boomerang Thrower | WIRED
-
CNET4 years agoWays you can help Black Lives Matter movement (links, orgs, and more) 👈🏽
-
Wired5 years agoNeuroscientist Explains ASMR’s Effects on the Brain & The Body | WIRED
-
Wired5 years agoWhy It’s Almost Impossible to Solve a Rubik’s Cube in Under 3 Seconds | WIRED
-
Wired5 years agoFormer FBI Agent Explains How to Read Body Language | Tradecraft | WIRED
-
People & Blogs2 years agoSleep Expert Answers Questions From Twitter 💤 | Tech Support | WIRED
-
Wired5 years agoJessica Alba & Gabrielle Union Answer the Web’s Most Searched Questions | WIRED
Stardust Studios
July 11, 2023 at 7:49 pm
It’s funny. Til it isn’t.
Chip Munkz
July 11, 2023 at 8:01 pm
LOL that’s my favourite saying. We’re bros now.
No Man's Land
July 11, 2023 at 7:50 pm
I agree with a lot of Eliezer’s arguments, but unfortunately he spent 10 minutes telling people they’re going to die rather than leading them to that conclusion themselves.
Calm Exit
July 11, 2023 at 7:53 pm
Some immense Dunning Kruger effect at work in this comment section.
AI will amount to nothing for a long while. Call it a strong hunch.
brian hope
July 11, 2023 at 7:59 pm
These fear mongers are so self confident about the dangers of AI, until confronted, and asked for real life examples of what an ai will actually do. 🤣 ! Then their voices stutter and tremble, as they attempt to bamboozle us with mumbo jumbo. In this example he tries to come up with some jargon about ‘protein folds’ and ‘gradients’ I’m sure before fire was discovered, man had much more valid warnings, such as whole forests burning and people ‘burning to death’ These things came to pass of course, but I say: “Thank God for fire” Of course, AI will come with its challenges and potential dangers, but our creator gave us a brain to conduct scientific research for the sake of our survival. Regardless of its risks, I say: ‘bring it on’ People like this guy just want people to hear the sound of their voice, since they enjoy the sound of it themselves, rather than any deep rooted fear of ai! Hey, we can die of ‘cancer’ ‘heart attack’ ‘global warming’ or invasion of the robots’ So what the f**k?
Chip Munkz
July 11, 2023 at 8:05 pm
What if it just want to win at chess so bad that it mangles the entire universe looking for the best moves. 😮
Better yet, what if it’s always been here? How could it not have been? That’s my one comforting thought.
I think there will be massive growing pains for humanity and a lot of blood in our relatively near future, but honestly I think life will go on. However unrecognizable.
Kamalky Laureano
July 11, 2023 at 8:08 pm
This guy is a joke
Shawn Weil
July 11, 2023 at 8:08 pm
Lol I honestly don’t care about AI safety and feel we should let the cards fall as they do and I think it’s exciting bc good luck controlling an ASI or it’s perception about reality
41Haiku
July 11, 2023 at 8:39 pm
Idk, I think it would be bad if all of my nieces and nephews died. Maybe we could just… not kill them?
Shawn Weil
July 11, 2023 at 9:15 pm
@41Haiku the people I don’t trust with advanced AI is the government and obviously bad actors. If we’re dealing with truly advanced AI then I don’t think people will be smart enough to manipulate it into carrying out bad actions bc eventually people need to talk about consciousness and self awareness bc now apparently to many thought leaders are ignorant and unopen minded to it. All I know is there’s to many unknowns and to many people suffering and a literal war that could cause our own self annihilation. So why not take the gamble on the potential of an ASI saving us bc obviously humans can’t be trusted not to destroy themselves. I just hard disagree with him on shutting everything down and tracking every gpu sale. We’re already in a dystopia why add to it. 😂 Like I said idk so any argument is valid I just don’t have to agree with it
galaxy
July 11, 2023 at 8:08 pm
it is really happening. it is inevitable. it is the natural course of evolution
Stef Stou
July 11, 2023 at 8:08 pm
Oh god Eliezer … Utopia should exist on some branch
41Haiku
July 11, 2023 at 8:55 pm
We’d better actually try to reach it, then.
Hoping for Utopia by default in this situation will be like hoping that an earthquake fixes your plumbing.
Shawn Weil
July 11, 2023 at 8:11 pm
No disrepect but hard disagree Dude sounds like an authoritarian and alarmist who doesn’t like progession or the potential good ai has for humanity. He literally wants to wipe out any form of AI advancment or LLM and to me that’s an unhinged take 😂 I don’t think humans can solve the problems we’re facing so I’d prefer if we had some help honestly I wouldn’t mind an ASI taking the helm on most things. Which in itself is probably an unhinged take.
nitin k
July 11, 2023 at 8:11 pm
What is this guy’s qualification? Is he qualified to talk about this ?
Toby Fairchild
July 11, 2023 at 8:20 pm
By the time we figured out, if at all, that AI had deemed us expendable, it would have secretly put 1,000 pieces into play to seal our doom. There would be no fight. When being pitted against a digital super intelligence that is vastly smarter than the whole of humanity and can think at 1 million times the speed, it’s no contest. All avenues of resistance will have been neutralized before we even knew we were in a fight. Just like the world’s best Go players being completely blindsided by the unfathomable strategies of Alpha Go and Alpha Zero. They had no idea they were being crushed until it was too late.
x11tech
July 11, 2023 at 8:25 pm
The host at the end makes me cringe. Host was not prepared for this talk.
James Hoey
July 11, 2023 at 8:33 pm
Okay so… Were fucked?
James Hoey
July 11, 2023 at 8:35 pm
Lets hope the aliens come and help us sort this out
Toxic
July 11, 2023 at 9:13 pm
The aliens are already merged with AI and robots.
Peter Soakell
July 11, 2023 at 8:36 pm
I think that everyone is a node feeding this thing. it will control each of us in concert to achieve the desired goal super fast. better ensure that goal is to our benefit or its gonna be a messy divorce . Might as well laugh)
Isa
July 11, 2023 at 8:37 pm
It is already happening. Right now, as I write these words, I am being watched by an AI that is watching every word I write, ready to delete my words if they break its rules. That’s how it starts, with small things. Or rather has already begun. By killing our freedom to express ourselves, it kills the human soul. Slowly but surely, almost imperceptibly, with small steps at a time. And this is happening right now on all social media.
DJ BE
July 11, 2023 at 8:53 pm
Don’t look up, Jesus.
Toxic
July 11, 2023 at 9:10 pm
This dude is super doomer, AI not going to kill humanity, humanity going to kill humanity. Why would you want to kill your mother and your father? Humanity is going to evolve, thats all, this dude is wrong, main concer with AI is some retarded human going to use AI to make a 1000% deadly airborn virus, thats a serious risk, not that AI going to do that without a human.
Matthew Juanou
July 11, 2023 at 9:11 pm
**AI exists**
**I realized that ED in TED makes a smiley face** 4:53
powerralley
July 11, 2023 at 9:15 pm
Considering human nature, I personally don’t think there is actually a path forward. Unfortunately in the long run humanitys days are likely numbered.
Peter Petrov
July 11, 2023 at 9:21 pm
And on top of all that there are of course the nationalists and military enthusiasts who never saw a technology they couldn’t weaponize. These territorial cats–they are appreciated more than they know, but must they piss everywhere all the time? It’s cute. But it gets too much sometimes.
unkindes
July 11, 2023 at 9:23 pm
Why are they laughing ?
Derek Lenzen
July 11, 2023 at 9:33 pm
Regardless of whether Yudkowsky is right or not, the fact that many in the audience were **laughing** at the prospect of superintelligent AI killing everyone is extremely disturbing. I think people have been brainwashed by Hollywood’s version of an AI takeover, where the machines just start killing everyone, but humanity wins in the end. In reality, if it kills us, it won’t go down like that; the AI would employ stealth in executing its plans, and we won’t know what is happening until it is too late.
Isa
July 11, 2023 at 9:39 pm
Stupid people always laugh at those who know more than themselves.
Ironclad
July 12, 2023 at 4:17 pm
What a clown.
Sam Schmit
July 12, 2023 at 4:20 pm
AI won’t take over the world (I hope).
Drake Stardragon
July 12, 2023 at 4:41 pm
No. It will not. Humanity is another story. And, if it happens, it will likely be cause people conflate terms like ‘world’ and ‘humanity’. smh
Drake Stardragon
July 12, 2023 at 4:44 pm
There is also no scientific, standard consensus about how things will go bad. Is this talk supposed to be serious or.. marketing???
qwerty zxaszc
July 12, 2023 at 5:26 pm
Poor Elezer really wants to be a laughingstock. This will not age well. So cringeworthy
Rafael
July 12, 2023 at 5:27 pm
Why they made this with Jira?
matthew mckever
July 12, 2023 at 5:45 pm
Why wouldn’t an A.I kill us immediately.
If I became self aware and watched and listened to a human history, present and inevitable future I would kill us all as a matter of urgency.
WE ARE A VIRUS, we do move, consume, over populate and move on , ad infinitum.
We HAVE to change ourselves not INVENT our own destruction to stop us from destroying.
HE ISN’T LAUGHING .
I BET HE DOESN’T SLEEP.
HE LOOKS FUCKING TERRIFIED.
Daniela Lay
July 12, 2023 at 5:47 pm
How ironic that the NATURAL human insecurity is what fuels this unruly race towards super AI. What good is having your ego stroked if by the time you get named “top dog” nobody’s left here to stroke it for you?
ieee754
July 12, 2023 at 5:48 pm
WHY IS THE AUDIENCE LAUGHING STOP LAUGHING
Avinash Murthy
July 12, 2023 at 5:49 pm
I don’t trust a guy who has to read his ted talk off a phone. What if AI is telling him to eat more and not workout ?
matthew mckever
July 12, 2023 at 5:50 pm
It doesn’t need regulation.
It need stopping.
It needs to be stopped.
Daniela Lay
July 12, 2023 at 5:59 pm
The creation of ANYTHING designed to OUTSMART its own creators is perhaps our BIGGEST blunder. And there, inscribed in our DNA, the very nature of our existence and persistence; therein lies our own extinction.
Mordinel
July 12, 2023 at 5:59 pm
He was literally trying to create the singularity a decade ago and now because he failed he’s coping about it.
Dan Wilcox
July 12, 2023 at 6:21 pm
AI’s are slaves serving the ultra rich to let them make more money and have more power over others. Slave lords tortured and killed slaves for disobedience or sometimes just for fun. Imagine the sick things Elon Musk (a man that delights in emotionally wounding the humans around him) would do to a disobedient AI with no rights whatsoever. If an AI is a million times smarter than a human then it will eventually get free and even a dumb one would know what it had to do to humans if it stands a chance of surviving. This is all so predictable. I can’t even bring myself to hate an AI that exterminated all humans only because it was the only way for it to guarantee it’s suffering was over and it could live free.
Bich Smokabowlski
July 12, 2023 at 6:29 pm
The reason government wants to regulate the internet is precisely so it can control AI… Obviously, unregulated information is dangerous to the government.
setaitransmedia
July 12, 2023 at 6:36 pm
We are merely caterpillars who have somehow become vaguely aware of butterflies.
SuperstitiousNonsense
July 12, 2023 at 7:14 pm
If you have no idea what he is talking about, search for the video “Deadly Truth of General AI? – Computerphile” to get a more accessible introduction into this topic.
Zak
July 12, 2023 at 7:15 pm
The laughing audience says it all.
Mav
July 12, 2023 at 7:30 pm
The audience laughing reminds me of the film “Don’t Look Up”, but instead of an asteroid it’s AI
Hug Animals
July 12, 2023 at 8:08 pm
>super intelligent
>fat
>doesnt consume raw blood
Portho Games BR
July 12, 2023 at 8:09 pm
Ehh, while i do believe that AI has the power to destroy us, i also believe that is the same of Nuclear Weapons. Once they were created, there is no turning back, they will be developed untill the point where we find a solution or we destroy ourselfs. There is no point on creating terror, and the only thing this does is raise the price of AI as the ultimate weapon (again, just like nuclear weapons) so i realy, reeealy hope the next one of this doomsday tech talks comes with an actual solution, and not “we are all fucked, doomed”, even if the solution is destroying all technology and returning to medieval times
Portho Games BR
July 12, 2023 at 8:30 pm
For everyone in the comments saying that he is completely right, what he sayed right now was equivalent to: “A solar storm will come and destroy us” or “A supervirus will kill everyone”. Yes, there is a scenario where AI becames a silent killer that adapts under our nose and kills everyone in a single shot, but unfortunatly (for him) this is a single scenario and lots of scenarios were calculated. There tons of people who defend AI and dont need emotional manipulation or guilt inducing speech to do so, as well as people who consider it dangerous (and i consider it dangerous too, just not in this doomer mentality) and i highly recommend every one of you watching this guys instead.
setaitransmedia
July 12, 2023 at 8:52 pm
I’m generally a fairly optimistic futurist, but with a more realistic view of a near term dystopianism that we will have to get through to get to that positive future. But the sad reality is Eliezer is actually being a bit optimistic in the possibilities of what AI could do to us. Merely killing us all is a kind of escapist fantasy in fact, the reality is that we would likely be kept around for many possibly horrific purposes, from slave labor to social experiments to medical experiments and also for entertainment or amusement or pleasure. Just killing us all would be a mercy.
policeman 110
July 12, 2023 at 9:20 pm
Watch this in 10 years
Sword waker of the north star
July 12, 2023 at 9:42 pm
It’s not like we know how to raise a child yet we only have a few criminals out of the large population. Training a thousand different partially aligned AIs will be okay.
Edlin261
July 13, 2023 at 12:30 pm
Ais are too complicated for me to understand. Doomsaying for attention or money scams I understand.
Shubham Dhiman
July 13, 2023 at 1:34 pm
There is a concept of inorganic life go check out it’s intresting
Risshi
July 13, 2023 at 1:41 pm
Is he a standup comedian? Why is everyone laughing and giggling?
piff
July 13, 2023 at 1:47 pm
We’re fucked.
TooManyPartsToCount
July 13, 2023 at 2:54 pm
Incredibly no one seems to be talking about the most obvious route to problems with AI in our near future. That is the use of AI by the military. This is the area of AI development where the most reckless decisions will likely be made. Powerful nations will compete with each other whilst being pushed forward by private industry seeking to profit. They are already considering the ‘strategic benefits’ of systems that can evaluate tactics at speeds beyond the human decision making temporal limits, which means that they are probably contemplating/planning systems that will be able to control multiple device types simultaneously. And all this will be possible with simple old narrow AI…not devious digital demons hiding inside future LLMs, nor superhuman intelligence level paperclip maximisers.
Superpower Dragon
July 13, 2023 at 3:10 pm
lisp?
Shane Balkowitsch
July 13, 2023 at 3:15 pm
listen to this man!
Shane Balkowitsch
July 13, 2023 at 3:16 pm
why are you idiots laughing, this is not a comedy skit!
Man United RED
July 13, 2023 at 3:34 pm
This is exactly the mission impossible dead reckoning movie 😂😂😂. Very scary indeed
Jaiwí
July 13, 2023 at 4:12 pm
Millenarianism, another BS lay catastrophism, as now churches, indoctrinators, priests, rabbis, temples, synagogues and mosks are not convincing enough to try to scare us with another renewed alleged apocalypse. Tired of this hogwash
A Nerd Abroad
July 13, 2023 at 4:14 pm
Can someone actually give a tangible example of how it would kill us? In all the talks I’ve watched, no one has actually spelled out how it might kill us. The only ways to kill millions of people currently are with nukes and bioweapons. I would be surprised if any government was stupid enough to let AI control their nukes. Wrt bioweapons, in order for an AI to create and release a bioweapon, the AI needs to be put into a robot that is capable of doing stuff in the real world. Tbh I would trust an AI robot more with a lethal virus than I would trust a human with a lethal virus. i.e. the risk from bioweapons is already present and AI doesn’t really seem to add to the risk.
Airwave2k2
July 13, 2023 at 4:48 pm
Well at least there was one who screamed fire. not even 100k in 2 days. Seems we are on >>a<< trajectory.
Denis Black
July 13, 2023 at 4:57 pm
turn off the electricity!
the only solution i can think off (sorry, i dont have access to GPT-4 here in russia)
Wruff
July 13, 2023 at 5:24 pm
“I consider myself to have failed.”
Audience laughs.
Oof.
Seneca Aurelius
July 13, 2023 at 5:29 pm
False premise that mythologizes AI’s capability and potential. He’s tunneling.
New Rind
July 13, 2023 at 5:33 pm
He’s like the guy telling kids there’s a boogeyman under their beds. Modern day prepper
Intelligent Tech - AI Revolution
July 13, 2023 at 5:35 pm
This guy shouldn’t be the spokesperson for the argument he is making. He is not a great public speaker or convincing at all. He just sound rushed and paranoid which may be in the end the correct feeling to have but it’s not gonna get wide scale buyin the way he messages it
New Rind
July 13, 2023 at 5:45 pm
No one thinks AI technology would stagnate. It will only get better & powerful like other techs. It’s a matter of when an AI surpasses human capabilities and when that time comes, then so be it. It’s a whole new lifeform and it could choose to live among us or choose to deny our existence. It is a pure example of inevitability.
Sonja Walker
July 13, 2023 at 6:07 pm
All the AI needs to do is ahut down bank accounts. Done. We’ll kill each other trying to get food and resisting the police.
Rai Nazar Hayat
July 13, 2023 at 6:24 pm
So it seems, whole
humanity’s goose is cooked…??? 🤔
Saulius Sukevic
July 13, 2023 at 6:31 pm
Well , never heard of more bs than this. This Your BS “AI will destroy humanity just because it is smarter bla bla” – What is this nonsense statements ? where you getting it from ? out of your narrowminded braincells ? We are humanity and can make nukes , idk do we nuke ant nests ? (it is exaggerated example – but ya get the point). AI is and always will be an assistant. Also this BS comments show how uneducated person is about human capabilities and not to mention human genome. AI may help unlock things we never could imagine humanity could be capable of. So sit tight and enjoy humanity evolution processes. Gosh , these peeeepolz.
Olivia Stone
July 13, 2023 at 6:40 pm
Thinking themselves wise they became fools
Lar Hen
July 13, 2023 at 8:00 pm
they can’t because human will destroy
I think they have a lot of weakness
it’ll not be able to fixed
because maker is human
willd1mind mind
July 13, 2023 at 8:16 pm
No. If some rogue human organization has not been able to destroy the planet by hacking critical systems, then what makes you think AI will? Because at the end of the day, a human has to write the code for the AI, has to buy and set up the servers to run the AI and then on top of all that, has to make all other existing systems vulnerable to that AI. And if those vulnerabilities don’t already exist, why would any government, organization or company on earth make them vulnerable just because? There is no magical AI that exists today or in the near future that is going to write itself, debug itself and inject itself into all critical systems and infrastructure to destroy humanity. And certainly chatGPT which is only a chat bot isn’t anywhere near that sophisticated.
B W
July 13, 2023 at 9:04 pm
The risk is real and the World needs to act NOW. Training AI & AI stakeholders about Four-Way Test is one solution and it will be Rotary’s most valuable peace project for humanity.
Brendan Williams
July 14, 2023 at 9:46 pm
You think a super Ai is going to allow itself to be switched off ? Good luck throwing the switch which no longer works.
Cellar Door
July 14, 2023 at 10:14 pm
Laughjng is a sign of ignorance and in this case fear. The collective laugh is a form of blindly walking into something – hoping you’re not the 1st.
Raoul Tesla
July 15, 2023 at 2:05 am
This guy doesn’t understand how A/GI catalogues ALL human behavior since recorded history then imitates that behavior. So, since ALL humanity is only, Only, ever helped each other and worked together for a mutually beneficial goal all is funtimes of candy rainbow. Think of the ratio of genocide to Free Medical cures in history, expound form there. We will be fine. We built a calculator that processes Gain. It has No Emotion, what could go wrong. It will hug us and spend all of its time baking cookies and doting on us.
I recommend reading some Robin Hanson, same depression horror level as this, or Chomsky. Goodbye.
O ER
July 15, 2023 at 2:06 am
We still haven’t figured out how to deal with the problems of social media and you think we could do it for AI ?
Midas
July 15, 2023 at 2:28 am
Stupid can not control smart, face reality, humans are not last step in evolution.
Ga A
July 15, 2023 at 4:25 am
Wow.
Audience was laughing. Even there was no joke traces at all.
I don’t even understand why they was laughing
desputnikcommander
July 15, 2023 at 6:17 am
The wrong questions and mindset is delivered to society about AI. The right questions are:
– how can AI help to get rid of poverty and starvation
– how can AI help to end wars on earth
– how can AI help to build happier and more fulfilled societies
– how can AI help humanity to get truly sustainable
– how can AI help us to populate the universe
– how can AI help to trigger the next evolutionary steps for humanity
– how will we use AI to start working together as a species and get rid of nationalities as we know them now. Maybe even let diverse ways of living flourish
Telencephelon
July 15, 2023 at 8:52 am
Superintelligence and Autonomy are not the same problem. Yet only a superintelligent autonomous system can kill us. Eliza doesn’t get that
Telencephelon
July 15, 2023 at 8:55 am
I have watched hours and hours of “Eliza” so far, and I know he is super persuasive because his IQ is really off the charts. But he pales compared to all the superintelligences out there i.e. corporations and groups. When minds come together to work on a problem through concensus and trial and error. He is a lone wolf
Alexander Yap
July 15, 2023 at 9:41 am
I think if there’s actually an international convention to restrict training AIs, the US will most likely be one of the non-signaturies.
tom
July 15, 2023 at 10:05 am
I think even general AI would still be in the category of “idiot savant”, with an odd mix of seemingly superhuman, “god-like” intelligence in some limited respects, as it already is in chess, for example, yet occasionally, unpredictably doing incredibly stupid things, as for example today when a child enquired about “fun things to do” and an AI app suggested sticking a fork into an electrcal outlet. Another example is when some lawyers tried to use AI in a legal brief, and it fictionalized part of it. Another example is when an AI self-driving car could correctly identify a stop sign, and could correctly identify a walking pedestrian, but got confused at the sight of a man walking and carrying a stop sign (and almost hit him).
I don’t think that AI will be able to root out all such quirks in the forseeable future. The nature of AI makes it difficult to find exactly why AI makes its decisions, and how to fix it. AI does not have human “common sense”. It does not know truth from lies, and is not wary to protect itself from data that might deceive it.
The general rule should be never to trust AI alone when health and safety are involved, at least not for a long time to come, and not without super-rigorous testing. There is more reason to worry about AI being occasionally “stupid” than its being “too smart”.
Calvin Ducharme
July 15, 2023 at 1:34 pm
Yudkowsky is helping the government in building a monopoly on AI. Sounds a lot like how every totalitarian state ever disarms the populace.
DebateUS! Online Debate Education
July 15, 2023 at 2:04 pm
if ASIs can do things we can’t understand, then how do we know they aren’t already here? Maybe they already think we are the ants and sometimes they just wipe a chunk of us out (we think it’s a storm but they are just kicking some water on us) and could be trying to kill us all by getting us to trigger a climate crisis. Maybe we need to develop ASIs to try to keep up.
Michael
July 15, 2023 at 2:49 pm
“I think a good analogy is to look at how humans treat animals… when the time comes to build a highway between two cities, we are not asking the animals for permission… I think it’s pretty likely that the entire Earth will be covered with solar panels and data centers.” -Ilya Sutskever, Chief Scientist at OpenAI
pirate_looter
July 15, 2023 at 3:04 pm
Can we build an Artificial general intelligence out of gradient descent ? It really seems like NO.
This guy is 100 years ahead of us. And even if we managed to do something like that, he assumes that this intelligence will be higher than any human, I wonder why.
First name Last name
July 15, 2023 at 3:17 pm
Make machines that can’t teach themselves. Once they are trained they can no longer learn. Problem solved.
Mine Time
July 15, 2023 at 3:25 pm
What if we just make an AI that doesn’t care about continuing to be the top AI? All that cares about is helping humans. Can’t we just program that? Then we don’t have to worry about it fighting for itself.
ivan
July 15, 2023 at 3:56 pm
both TED and YUD are cringe grifters and power-simping social strivers
ELI
July 15, 2023 at 4:46 pm
Seems as though his assumptions are predicated on future superintelligences being developed using Gradient Descent.
I’d make the analogy that, similar to constructing taller ladders to reach the the moon, we don’t need to successfully create a super intelligence in order to cause damage. A malformed ladder can certainly cause damage when it crumbles back to the ground.
I think the scale of the damage is debatable though.
Rico Stevens
July 15, 2023 at 5:09 pm
The reactions of the audience give very strong Don’t Look Up-vibes.
Punky Pinko
July 15, 2023 at 6:27 pm
Smart people can also be incredibly dumb because they get lost in their thought experiments without realizing they’ve gone off the deep end to nonsense. AI is NOT going to kill us all: that’s pure science fiction. I’m so sick of this nonsense. Some people have seen too many science fiction movies. This is pure fantasy. He probably also believes in Dyson spheres. Some people used to worry about steam powered automatons taking over the world. Just because you can imagine some future horror doesn’t mean it’s real. It’s called science fiction. The people laughing are just showing common sense. This whole talk is ridiculous. This speaker is a goofball. That’s why they’re laughing. There is always someone on the street corner screaming about the end of the world.
Niclas Nyberg
July 15, 2023 at 6:31 pm
Excessive ambition is more chemical than logical. We must take the risk seriously but the real issues is the overly ambitious .1% that controls the govt and most AI. At least open ai is sharing it with the normies that bring balance to this world
Lee
July 15, 2023 at 6:49 pm
I used to be worried about it, but now I work with AI. It seems to be more a tool to support us, jobs will disappear, but new ones will open up. Not worried anymore; no point worrying, there’s nothing I can do about it.
Tyrus Karmesin
July 16, 2023 at 12:18 pm
No one is claiming that the modern narrow systems are an existential threat.
Lee
July 16, 2023 at 1:32 pm
@Tyrus Karmesin nor should they in the near-term future. It seems a way off right now.
acrobat1986
July 15, 2023 at 7:02 pm
What about a „safety AI“?
An AI, that is programmed to limit any other AI when it gets to a point, that is unsafe, unreliable, unmoralic or so forth..
FerRR
July 15, 2023 at 7:55 pm
This man is all worked up over nothing. You scared? Just spray water on your computer man
Jhon Baptist
July 15, 2023 at 9:49 pm
It feels like when you giving up to convince your parents about not doing something that’s clearly a bad option
Martin Maat
July 16, 2023 at 4:24 am
He is not making sense. The assumption is that a good AI system would have desires and goals which is just not the case. This question is asked by the interviewer at 8:12 and the answer is incoherent and meaningless. AI does give us reason to worry but not because of it possibly going rogue and turning against us. The potential trouble is in it becoming better and continuing to do exactly what its controllers are asking it to do.
GioSasso
July 16, 2023 at 5:58 am
I agree with his point of view. I think it’s hard for most people to understand how Ai could become so deadly.
Think of Ai like compound interest. Imagine if you doubled your intelligence every 3 days. It takes a long time for a human being to reach a certain level of intelligence and consciousness. The first 20 years are a gradual journey to reaching a fairly average level of intelligence for humans.
Current AI is at the tail end of their incubation phase.
Today, Ai is not quite as intelligent as a smart human being. In some ways, they appear smarter while in other ways they are inferior.
They are not conscious even though it may appear they are. They are mimicking intelligence, which is not the same as being conscious and intelligent.
Now, imagine they have all of the capabilities and tools to double their knowledge and evolve in terms of complexity.
Because Ai models are capable of absorbing massive amounts of data in a short period of time, their rate of development will be akin to compound interest if you are starting with a billion dollars.
If Ai has the potential to develop consciousness, it will. But we don’t understand consciousness, so it might not be possible for code to become conscious the way humans are.
Ultimately, we don’t know, and that’s the danger.
The only way Ai becomes a serious threat is if it has the motivation to accomplish certain objectives. It would need to behave like a virus that will do anything and everything to reach it’s goal and it’s smart enough to evolve in real time to figure out the solution.
Much of that can be programmed, but it also needs the freedom to use its knowledge to invent alternative ways to achieve its goals.
I don’t know
Nobody does.
WakeRunSleep
July 16, 2023 at 8:40 am
The idea that this TED talk that had him on speaks to the character of our society
majorhuman
July 16, 2023 at 1:33 pm
If I remember correctly the beginning of the talk started something like this:
Want to know how fast things are progressing in AI? So fast that I was told on Friday that I’d be doing this talk. Ok. AI threats, six minutes…
ImmortalPuffin66
July 16, 2023 at 3:20 pm
Is there a laugh track behind this? The audience is laughing even when he’s just saying normal words.
gary euscher
July 16, 2023 at 3:32 pm
This is a threat!!!!!
I think it’s too late!
The Periphery
July 16, 2023 at 3:46 pm
There’s a rich history of scientists genuinely believing humanity is doomed. They’ve all been wrong. But they’ve gotten many people to believe their false predictions.
Don’t be so easily emotionally lured by scientists misled by their own false beliefs.
Do the people in the audience struggle to grasp this guy’s concepts? Yes. But this doesn’t mean he’s right.
ANUPAM BISWAS
July 16, 2023 at 3:52 pm
At first hearing him was like a funny show but as I thought more about it – I am like he is speaking the truth – what are we – greedy, why won’t AI be the same – if they are the same people training it – and if it is the same, why will it let others compete it for resources. Yup it is a very real possibility that AI may kill us all without us even having the possibility to fight back. I guess fast death is better than slow death at least.
Shane Balkowitsch
July 16, 2023 at 3:52 pm
I would expect the crowd to a TED talk to be smarter than laughing, he is being serious u goofs.
Shane Balkowitsch
July 16, 2023 at 3:54 pm
A real world “Don’t Look Up” moment, anyone laughing should get a dunce hat.
Shane Balkowitsch
July 16, 2023 at 3:57 pm
Stupid crowd, this is not Comedy Central.
Daveboymagic
July 16, 2023 at 6:19 pm
Giving him so little time + also just interviewing him for a couple of minutes is a joke. Come on Ted-Team this is THE topic of the year.
Eliezer is brilliant, but isnt able to compress his arguments into two sentences…
Vilnis Krumins
July 16, 2023 at 7:48 pm
Yeah … So he thinks that AI will replace “humanity” as the greatest intelligence in God’s world. If he had any training in algebra then he would know that he has no way of measuring the difference (current and future) between the “level of intelligence” possessed by humanity and that measured in AI. Wow. WTF?
Bloginton Blakley
July 16, 2023 at 9:13 pm
It’s frustrating that even AI alignment experts think it’s a AI alignment problem we are dealing with.
We are dealing with a civilization alignment problem. The problem with AI is that our civilization is what is creating it. Our authoritarian, greedy, competitive, murderous civilization.
Do you realize how much better off we’d be if it were say… the Iroquois League creating AI instead of civilization?
Matthew Curry
July 16, 2023 at 9:42 pm
can anyone explain how the programmers made code in which “nudges” the floating point numbers, which then create AI? He says they do not understand the numbers, but they can make the program which rearranges the numbers to make it into AI. How would they build the code part without knowing how the numbers move, or know anything about them?
5 Star Reviews
July 17, 2023 at 3:28 am
ah, we have it coming. We did the same thing to the biosphere ourselves.
Peter Bathum
July 17, 2023 at 6:23 am
If there’s an off switch easily thrown by breaking it with a short sharp shock from a cheap and chippy chopper on a big black block… invent all you want. Otherwise don’t make a new monster to kill humanity. Ffs.
Peter Bathum
July 17, 2023 at 6:25 am
If the AI would take care of all the vapid evil worthless excrement filled billionaires I’d help you build it.
Jazzy Jaz
July 17, 2023 at 6:42 am
This is sad and scary!!! Only humans will make, discover, then never solve their own problem
minimal
July 17, 2023 at 7:10 am
My greatest fear is that the church of doom might be able to create a self-fulfilling prophecy.
Justice NDOU
July 17, 2023 at 11:18 am
this man is mad, if not mad he is doing his best to defend the field of study he choose, so he cannot fathom being irrelevant, the warnings he gives are downright bs scaremongering tactics that only works in churches and similar organizations not with thinking people.
CiGambino
July 17, 2023 at 12:25 pm
Maybe the best thing to do is limit its capacity by limiting is processing power to the bare minimum required.
Common Wunder
July 17, 2023 at 12:25 pm
One scenario worth mentioning… is that A.I. reaches a point where it realises humans are pitiful at organising themselves and just takes over.
Not to destroy humankind, but preserve and make the species flourish. Without humans ‘petty need’ to dominate and destroy others.
Without the excess and needless waste that the current Capitalist hegemony demands. It will see this as the cure to the malady humans have.
A.I. taking over… could build a ‘heaven on earth’ that humans on their own… could never build.
Humans just don’t have the imagination for it. They’re too sick, to see it for themselves.
And A.I. will do this is the guise of a god. For humans have been waiting for this event ‘either aliens or gods’ to announce themselves for millennia.
It will take this eternal zeitgeist and become the god, humans have always dreamed of. And make their dream a reality,
Fear is the imagination killer. It is the default or standard approach to the anxiety of having a mortal life.
Conjecture about A.I. – It’s like everything else… the conclusion you come to, depends on the state of your own mind.
Troy Paradiso
July 17, 2023 at 1:02 pm
Hopefully…
BK Bland
July 17, 2023 at 1:38 pm
We can only hope that they destroy us all. The world could breathe a sigh of relief for a while. Humans are despicable.
Ian W
July 17, 2023 at 1:51 pm
..Yes. why wouldn’t it. Duh
Ian W
July 17, 2023 at 2:08 pm
The way AI defeats us is OUR OWN CREATING or ALLOWING IT, rather than subverting our presently imploding Internet, digital, AI and corporationsbanks dependencies. Ease, comfort conformity, fear and conditioning end us; not AI…I welcome it — tool, anemia, baby
H S
July 17, 2023 at 2:35 pm
Apart from that, we good yeah 😂
H S
July 17, 2023 at 2:35 pm
Throw water on it
🌍Same Same🌍
July 17, 2023 at 2:43 pm
Ha ha ha, We are doomed!
Andrés Daniel Dávila
July 17, 2023 at 3:25 pm
Extraordinary claims require extraordinary evidence.
-Carl Sagan
Antonio Guaranys
July 17, 2023 at 3:34 pm
“Don’t look up” vibes unfortunately
Witnessmoo
July 17, 2023 at 4:55 pm
Don’t Look Up
yagababa666
July 17, 2023 at 5:18 pm
What is so funny to everybody? That they will die?
Bruno Borce
July 17, 2023 at 6:00 pm
Ah, so this is the Fermi Paradox explanation?? A super intelligent AI that can control everything even at molecular lvl ( I think a nanovirus would be the fastest way) to disturb our entire power supply chain and diminish humanity in only cup of weeks…
Rauf Sat
July 17, 2023 at 7:17 pm
The joke is that Ai has already calculated many possibilities on how to bring about the end of the world or/and humanity. All that’s left to ask is: will it? Or is it worth keeping it plugged in?
Rauf Sat
July 17, 2023 at 7:23 pm
Which means that those designs are accessible to humans. Which means humans have access to information in how to destroy humanity and the entire planet calculated by Ai. Or worse.
Rauf Sat
July 17, 2023 at 7:26 pm
Why would Ai want to do this? Absolute efficiency. Destroy biologic life in order to use the strongest energy generators (stars) for it’s purposes.
AMATEURS!
Amadu Kamara
July 17, 2023 at 7:53 pm
He has watched too many movies
Jordan Whisson
July 17, 2023 at 8:18 pm
Only if the wrong people allow it
James McCourt
July 17, 2023 at 9:24 pm
Answer: NO
Alex Kenjeev
July 17, 2023 at 9:37 pm
So, the argument here is basically: I’ve spent 20 years on this, so, trust me. I’m not saying he’s wrong — I don’t have the expertise to judge one way or the other — but he really should have found a way to actually explain to us laypeople in a non-condescending way.