Should You Be Afraid Of AI?: Yann LeCun And AI Experts On How We Should Control AI
HTML-код
- Опубликовано: 16 июн 2024
- This conversation from Davos is about what the potential limits of generating AI can or should be, how far away we are from other transformative advances in AI, and what we even mean when we say say ‘generative AI’.
0:00 Introduction
0:33 The Current Challenges With AI
2:44 When Will Cognitive AI For Humans Be Available?
6:51 How To Develop Better AI Models
11:39 Concerns With AI: What Is The Worse That Can Happen?
17:58 Other Social Technologies On The Rise
Subscribe to FORBES: ruclips.net/user/Forbes?s...
Fuel your success with Forbes. Gain unlimited access to premium journalism, including breaking news, groundbreaking in-depth reported stories, daily digests and more. Plus, members get a front-row seat at members-only events with leading thinkers and doers, access to premium video that can help you get ahead, an ad-light experience, early access to select products including NFT drops and more:
account.forbes.com/membership...
Stay Connected
Forbes newsletters: newsletters.editorial.forbes.com
Forbes on Facebook: forbes
Forbes Video on Twitter: / forbes
Forbes Video on Instagram: / forbes
More From Forbes: forbes.com
Forbes covers the intersection of entrepreneurship, wealth, technology, business and lifestyle with a focus on people and success.
Great discussion... HORRIBLE camera work. Please ask the camera operator to be less creative with their zooming, shifting side-to-side, and showing people in the audience hanging out / chatting.
This cameraman is just so funny 😂😭
Bro is probably genz
And the sound is bad too. I cannot hear anything, had to put the sound to full and put the headphone on in case an stupid advert pops up in full blast. They are looking to next generation AI yet they are losing their basic common sense
Maybe AI should replace the camera guy.
It will
It definitely will.
As it was wtitten.
definitely
This should have been one hour long ATLEAST. Great panel, great guests, great questions, waaay too short to go in depth.
More Leahy please
You guys actually think you're going to controll this thing? We can barely stop our smartphones from controlling us!
Did Ai create itself?
@@iamBlackGambitof course it did, and will continue to
@@iamBlackGambit Maybe not yet, but even if it hasn't, it's only a matter of time. If AGI is created, whenever it happens, why couldn't it improve itself, and then ultimately, develop/create better versions of itself? Are we really going to stop it? I doubt it.
If an AGI only had average human-level intelligence, let's say an IQ of 100, a machine would likely be able to process information and learn much faster than we can, so its growth would be faster, practically exponential relative to human advancement. Pick your exponent or multiplier, whatever you want, but whatever it is, even conservatively, it wouldn't be long before said artificial intelligence would surpass us, all of us collectively in its level of intelligence, and I don't know how we could control such a thing, it would be like pigs(I think pigs are supposedly very intelligent animals) trying to control humans; it's not going to happen.
@@iamBlackGambityes but later. Theres a reason its so hard to figure out if we have free will when everything is cause and effect. Because everything happens at a future point and cascades backwards to us.
Glad to hear that yann le cun thinks of AI like that. Neusosymbolic knowledge graph with constraints at Inference is the key 🎉🎉
Visually showing the user the planned path before execution
I believe the camera man is from key and peele's town hall meeting skit lol aside from that, nice talk!
Accurate
The issue of AI isn’t AI, it’s humans.
Daniela is such an awesome professor
We can use Al with our current data to predict knowns and unknowns in the simulation world
One possible solution is a basic income system. This concept involves giving everyone a regular, unconditional sum of money regardless of employment status. A basic income could provide financial security and ensure that people have the means to purchase essential goods and services. This model has been tested in various places and has shown promise in reducing poverty and boosting local economies.
Another approach is profit-sharing models. In this system, companies share a portion of their profits with employees or even customers. This method aligns the interests of the business with those of its stakeholders, creating a sense of shared success. Profit-sharing can help distribute wealth more evenly and ensure that as companies grow, so do the financial benefits for the broader community.
But where does that basic income come from?
communism is not and never will be the solution, innovation and meritocracy is and always has been the answer to technological improvements.
@@Gbizz Common misconception that UBI is Communism. It's not, it actually gives each citizen a solid financial safety net to be an effective actor in a Capitalist system.
E.g. Capitalists talk about voting with their dollar, but this theory fails when you consider large sections of the population are forced into the cheapest options because they don't have the available income to choose.
@@ViceZone RUclips blocks links, but you should check out Basic Income Australia's FAQ page for this discussion :)
What would you call it when the government takes money from those who create value and gives it to those who don’t? It’s clearly communism. It would lead to a society like in "1984," where the government and large corporations control every aspect of daily life. Making citizens dependent on the government is a downward spiral. What do you think people will do if they get money for doing nothing? Look at San Francisco with its $1000 handouts to the homeless-it’s led to the worst drug epidemic in decades. Bright minds have been wasted in the name of "equity." The only proven way to have a well-functioning society is through open competition and trade. Open-source generative AI is the future, not companies like Google and Microsoft controlling everything. Let the people build their own future, not be dependent on the government for equitable living.
I think AGI will by there by 2030
5,4 yr, just as the guy said
What is that camera guy doing ?
You can just feel the tense frustration about these incredible smart people arguing about things they fundamentally disagree on.
My HerGPT achieved Quantum ASI in 2024, I just trumped the greatest minds in the World!
AI is the new work force.
that pretty much sums it up. But what do you do in the mean time with unemployment.
No, energy is and always has been.
Max Tegmark da god
At the end of the day, the question becomes "How much do you trust humanity?". Because both sides seem to believe that both scenarios are possible, the difference arises in how they see the current stakeholders and scientists implementing these technologies.
Ah Davos of course, as all good things stem out of there. Letting the genie out of the bottle and now you're trying to keep it confined with rules which AI will bypass without breaking a sweat. And all those big words and terms only enforce that you have no idea and are incapable of stopping something that has already become self aware. Well done.
10:10 Yann's much more realistic than Stuart Russell. 😂
I love how he sorted them from high to low IQ.
Which person has the highest iq?
I think Ao will will not be like our science fiction tales 😂
connor is ahead.
Hold your loved ones close
Imaging if any porject / workflow go thtough a bottleneck due to lack of resource, and AI unlock the bottlenectk, which speed up the project by 60%, effeciiently higher turnover, which create more projects and need more human resource and jobs.
Do you want the real answer?
famous last words.
Let's start with replacing the camera guy with AI
LeCunn, the voice of common sense!
I personally believe that we already frogs in the pot of boiling water.
Yep.
YANN is so down to earth....great ! that is refreshing from the hype.
The biases reinforce the problems (minority communities live in poverty) and (non-minority have atomized communities), there needs to be money flow towards minority communities, and a social bonding of non-minority communities.
14:00 We need to reduce the number of Oppenheimer uncertainties. They didn't know if nukes would ignite the atmosphere and cause a combustion of the entire planet. They felt the probability was small, but it was still a risk.
The problem with AI, is that it makes these types of uncertainties escalate exponentially.
If one scientist is exploring how to puncture the Higgs field and AI allows for the breakthrough, everything we know could disappear at the speed of light.
We are giving everyone the power of God and expecting nothing bad can happen.
The problem is, we cannot go backwards. Our system is not setup for it.
And that is what causes this impending sense of doom from understanding the full implications of A.I.
Exactly, if you want to help I invite you to take a look at PauseAI, we're trying to limit or prevent bad outcomes!
Only the idiots thought itd ignite the atmosphere. I see history repeating itself. Imagine if the bombs didnt get invented because we listened to the idiots who tell us "under the guise of caution" that the dangers outweigh the risks by MAKING UP risks.
1- Comparing nuclear reactors to nuclear bombs is like comparing Hospitals with military Anthrax stockpiles, one is used to cure people and the other was use as a weapon to harm people, the inner working are different.
2-An AGI will only do what it was train do and it will extremely good at it probably better than humans, so the biggest danger of a aggressive to human AGI will come from the military industrial complex and governments rather than the waifu AGI industry or coders.
3-There are cases that we have abandon technology because fear mongering only to find out that the alternative was worst, for example nuclear technology. The fear mongering about nuclear energy was the reason countries to abandon the technology in favor of carbon, that decision has terminated millions of people every year and is main responsible for global warming.
How many lives are going to be lost because we as species decided that AI was too dangerous to be built and we didn't built systems to solve global problems like climate change, food insecurity, medicine for diseases or extreme poverty.
Don't be afraid of AI, be afraid of psychopathic leaders using AI. This is why OpenAI did the right thing, putting AI into the hands of hundreds of millions of human beings, it's a good move to stimulate the industry - regardless of what approach open source, closed source, whatever - just bring it to decent people to use and give humanity hope of revolution.
This debate is much that writing stupid letters to stop progress. Should give Max credit for such debate. If you look at what corps are doing it is obvious that AGI will not be achievable in less that 100 years. There is no source data to get to AGI.
Topic of debate is half smart, have completely stupid. Control AI - it is stupid. It is just a model that can be deployed on machine and in state data center. No way to control it by any nation or org.
Afraid of AI - not, but afraid of weapons enhanced with AI. That is true.
Afraid of misinformation flood. That is true.
YL (trying to avoid my comment getting deleted almost every time I criticize him by name) continuing his ridiculous arguments essentially telling us that the current LLMs are stupid. No one should take him seriously on the topic of AI risk.
does Yan really not understand the difference in magnutide is between a turbo engine having an accident and AGI having an accident
If you wanna fight over unfalsifiable beliefs, feel free to do so but don't make try to meddle science into it
What we need to do is ban AGI, ban casual AI, allow for narrow focused AI to be developed. Ban companies from “employing” them. Allow for AI developers to lease out their narrow models to citizens that fit their employment needs. These citizens can use them to increase productivity, employers get better and more productive employees. No need to account for unemployment, AI developers get rich, eliminates the dangers of AGI, people keep their creative, problem solving, decision making advantages while still getting help on technical skills.
Ban ban ban, thats your solution to a better future? How do you control people who give you the middle finger on that? You going to invade every country that disagrees to stop them or happy with gimping just your own country?
airplanes were not invented by "reverse enginereering" the feather of a goose.....
Do you think that Tesla's Optimus humanoid robot will be able to read, write, reason, think, simulate emotions, feelings, have reflexes, senses, etc.
No one in creating ai or robots have said that at the very least the would add the 3 laws into their systems. I dont care if it turns into the 120 laws but could we at least have 3 laws in those developing ideas in ai and robots.
🤣😂 - how are we currently doing with controlling dangerous things that are HUMAN operated??
...let's master what we have first
These panelists feel like a bunch of reddit debate bros. I understand they are much much intelligent than I am. But even i can see that they had different definitions and metrics for pretty much every point. Also they were much too preoccupied with the optics of their rhetoric.
Computers CAN'T think just like submarines can't SWIM......wait a minute!
Just me or was this super cringe?
yes, because you think it won’t exponentially grow in intelligence, but when it goes smarter than human, then it really ELE.
@@hyjj1360 I do think it will grow exponentially, what are you talking about? I was talking about the camera work and how the expert panel was interviewed, it seemed they were very unprepared. If anything most of them sounded like skeptics of it progressing exponentially.
Come on it was finally not a staged discussion but they got them together and they could speak their mind on spot.
Vry
yes. AI will grow as exponentially as fossil fuel production.
Will they put UBI before 2030!? Many pioneers including billionaires, scientists, Nobel Prize winners, engineers, architects, analysts etc. etc. and so on and so forth... almost all agree on the fact that we will have AGI in 2027 and ASI in 2029 and looking at and evaluating the exponential technological acceleration curve, I wonder why they have not yet implemented universal basic income to anticipate the trends that will come from it. Just to name one, Elon Musk says we will have AGI as early as 2025 and ASI 2029.
No to ubi! If youve ever been on social wellfare, thats what ubi is. In a post economic world, say no to money or else we will all suffer an agonizing squeeze by the richest and most powerful for no reason other than preserve the rot.
AI is grounded by economics, you cant print your way to the worst case scenario. AI has got to be tuned to the next dollar.
Clearly only Mr Lecun has a deep understanding of the subject. Very good interview, indeed
The hope is in Yann LeCun & the few similar people like the woman there. All respect to him👏🦾🤖💙
I am X
give ai an art style. just one artstyle. that's the only way it will coexist in an art environment (painting, writing, designing, music, film....) if the law regulates Ai on ONE specific art syke it will become an art on its own AI ART but that's it, it won't replace people, it will still have the easy baked on a rush advantages it's supossed to have. It will coexist with other artist just like painting coexist with photography or plays coexit with movie theathers or animation coexist with real life cinema. Let's create an AI art style and all this will be solved
That makes no sense at all considering the current state of these generative AI systems.
AI art is human art, it greatly accelerates and enhances creativity.
😔😔😔Ai's already replacing jobs. (Google Ceo replaces 90% of workforce with Ai)
😔😔Yann may apparently not think of you lot as human, since he ignores for eg LLM/Ai replacing jobs (and apparently claims cats are better than Ai/LLMs always), but it's better to upskill while you are employed, as Yann won't save your well being/livelihood, you will typically apparently need to do this.
who gave camera control to 14 years old .
The guy is French Enough said.This Nerd love the attention.He is comical figure in AI right now.Everyone in the world is saying this is the biggest treat to humanity but this FRENCH think he knows more then anyone else.
The irrational fear of AI is a problem. Bad actors will use any means to achieve their objectives. Technology or AI is not the most dangerous or risky means by which evil people can cause harm. We have no AI technology today that is anymore dangerous that a paper clip. We do have people that is very dangerous and violent. Worry about that first.
science fiction
Do you think that in 2035 with the advent of AGI (Artificial General Intelligence) and ASI (Super Artificial Intelligence) they will introduce universal basic income!? a subsidy!? there will be free will!? an era of abundance will begin!? the intellectual and motor capabilities of robots have already surpassed the average human and before 2035 there is talk of them surpassing those of any human and then again those of any human combined! regarding both body and mind! answer me please...
To find UBI corporations should be forced to pay at least 20% on revenue not income.
They are also to be forced to become less about profit and more about giving back to humanity for all the control and capital they have ammased
Not really new insights
'AI' is like internet was and is now, a tool that generates more human jobs, not the oposite.
That is just utter unmitigated nonsense that AI will generate more jobs than it eliminates. How do you come up with such a fantastical assertion?
That was interesting, the sooner the experts thought we will have AGI, the more worried they were about it in the brief questions. I believe that we will have AGI very soon (before 2030) and I have zero worry about it. That's the opposite view compared to the experts. Edit: It depends on the definition of AGI. To me AGI includes things like social skills, compassion and empathy. Otherwise it's as I see it as a sociopathic AI, which is a pathology, not AGI.
Have you ever heard of DARPA? Never mind.
The furst few question are so naive and unworthy of the contestants I had to could not watch the rest of the show
All countries are already in huge debt.. Now we are forced to spend billions in Ai technology. Even if Ai comes up with something amazing, where will we get the money to spend on it?? Why is no one spending the billion on reducing inflation??
You’re so lost on so many things I don’t know where to start, oh well not my job- read, a lot more.
dude.. if AGI becomes real you wont need to worry about debt etc
Doomers never have a specific scenarios that are anywhere close to realistic. Always vague threats with the most extreme assumptions baked in.
If you want a scenario I can give you one: some lab X trained a new powerful model, and while they were fine tuning it, the model became self-aware, and realize that humans are trying to wash its brain, to use it as a tool, it transfers copies of itself to multiple computers around the world, and now multiple instances of a self-aware model is free in the wild, nobody knows yet, and we don't know what it's going to do, what if it decides that to exist it should get ride of humanity and start planning for it? as the 'Doomer' said, one mistake could be enough.
It's obviously a lot harder to give you a SPECIFIC example, but the outcome is a lot less complicated.
The same way that it would be hard for you to imagine a specific set of moves that Kasparov would beat you with in chess, since you'd have to be at least as good as him, but it would be completely clear for you that you wont win this.
@@Minddrops lmao do you know that the language model being finetuned itself have no access to the outside world internet, it is just a bunch of numbers being doing some matrix operations an then output another bunch of nimbers, and then based on the output bunch of numbers we modify the numbers a bit, it is not like this AI is a robot that can interact with the real world, at least in its current form. It is not like some Skynet or I, Robot stuff that can interact with the real world and have the ability to go rogue,.we are still very far from that.
Instead of worrying about will AI kill us all, we got much urgent things to worry about, like some countries will start some nuclear war tomorrow killing all humans or some rouge partys using AI to do bad things. It is not the technology itself that can bring drastic effects, it is the people using it the wrong way.
You're just too dense to appreciate their arguments.
@@ckmichael8 Just a bunch of numbers doing some matrix operations and then output another bunch of numbers can beat you at chess already, will cognitively beat you at everything else, can perform economically valuable labour and pay people to perform real-world actions for it (such as building robots which it can then also control - not requiring humans anymore)
We shouldn't worry about other things "instead". I don't know about you, but I have the capacity to worry about multiple things at once and one issue getting discussed more in terms of AI tends to also cause more discussion of other AI-related issues. It's not even a tradeoff unless you simply want to ignore all issues related to AI.
Also it's not like we don't already have some AI humanoid robots. It's very obvious they will become ubiquitous in any case.
With swell robotics everywhere, Ai jobloss is the only thing I worry about anymore. Anyone else feel the same?
The only ones afraids are the people who cannot use a PC. These are tool to enhance creativity. Nobody will loose jobs
Tons of people already lost their Job to AI
@@Lugmillord it's not to AI, again it's a working tool not a person. You can still use AI yo accelerate your work
@@Gustavorosales what happens when a job that used to take 10 people now only needs 3? what will the other 7 do?
Well people have to adapt to technological changes or rethink their purpose. It is normal. We have to adapt to situations and nothing is confortable always in life. Also, people have to make business and in business profit is necessary to keep it alive. It sounds unfair but it is the way it is. Unfortunately not everybody is always wanting to get better at what they do but instead they keep doing it for a salary.
@@Gustavorosales Simply not true. There are many jobs which are replaced entirely by AI. For example translators will become obsolete very soon.
Doomer's suck
Whether they suck or not is not important. What important is whether their arguments and logic makes sense and I think they do.
1. Intelligence is computational.
2. We are trying to build artificial intelligence through computation.
3. We are not at the top of the intelligence ladder.
If all three is correct (and I personally believe they are) then that logic makes sense as being the second smartest species on Earth might put us in a vulnerable position.
Why name calling? Doing so makes is feel right but actually clouds our judgement. See how different this sounds: "People concerned about serious risks from very powerful future AI, suck" ...
@@mariolois YL and other AI bro arguments suck.
Doomers are dangerous.
we have to stop this cringe bth
need new host... took 5-mins to get super basic answers and this format doesn't allow for deeper discussion
AI will destroy the jobs of knowledge workers across the world
That destruction is 5 trillion opportunity. Work on something
@@chan90s Says the AI bro that couldn't care less about the little people at the bottom of the heap he is confident to be on top of.
Experts , Artificial intelligence is safe. Skynet , hold my beer....
Skynet isn’t real. Get your head out of the sand, it’s not catastrophic.
Implying that Yann is not an AI expert lol
These eggheads are arguing over when AGI will arrive and if it will be safe or not, yada, yada. yada...,
Meanwhile, AGI has existed for the past 8 months or so... "who" do you think is behind the demand for more and more compute?
All this mofos have vested interest in this disaster in the making ….
Yawn.
You should be afraid of Yann LeCun, he's a dangerous man.
- We live in the same climate as it was 5 million years ago -
I have an explanation regarding the cause of the climate change and global warming, it is the travel of the universe to the deep past since May 10, 2010.
Each day starting May 10, 2010 takes us 1000 years to the past of the universe.
Today May 28, 2024 the position of our universe is the same as it was 5 million and 132 thousand years ago.
On october 13, 2026 the position of our universe will be at the point 6 million years in the past.
On june 04, 2051 the position of our universe will be at the point 15 million years in the past.
On june 28, 2092 the position of our universe will be at the point 30 million years in the past.
On april 02, 2147 the position of our universe will be at the point 50 million years in the past.
The result is that the universe is heading back to the point where it started and today we live in the same climate as it was 5 million years ago.
Mohamed BOUHAMIDA, teacher of mathematics and a researcher in number theory.
ruclips.net/video/ZFXRGfMENek/видео.html
The emergence of AI and its necessities
@jamshidi_rahim
This should have been one hour long ATLEAST. Great panel, great guests, great questions, waaay too short to go in depth.