I get what you're trying to say that at this point we can't put AI back into the Genie bottle which is so true. What I say is screw it it is a very useful tool now we just have to figure out how to harness it and try to stop it from being used by Bad actors and do what we can
@@Timely-ud4rm thats easy make ai a follower of Jezus and abide to Gods laws not a house build om loose sand! Problem is many AI devs are atheists or worse ;)
@@mpwebwerk What are you on about? AI is a tool to be used and harnessed. Thats like saying you want your toaster to follow the laws of god it's ridulious. Keep spiritual stuff humans and humans only and leave AI out of it what's next your gonna try to convince your shovel that it should believe in god? I am spiritual myself but AI isn't even alive or conscious for that matter it's just a sorting algorisiom that been trained on data other than what videos to recommend and what is the genome.
As a non US citizen it would concern me more if only the US had capable models. We need all the countries in world to be participating so we strike a balance. Concentrated power is not good.
A robot can be had for a few thousand dollars. An automatic weapon for a few thousand more. The weapon needs no software. The robot is a networked device with edge computing capability. A good programmer can create the lethal combination. Treaties will do little to stop this from happening.
Lol that's what worries the hegemonic status quo rulers - the fact that they aren't the ones with outright tech superiority anymore. Notice the whole "China is a problem" mantra only started being parroted in their controlled MSM when the Chingiez figured out how to play the economics & tech game and started competing there. No one ever complained about "over capacity" before as that was left up to the market to correct lol
@@frederickwelsh treaties don't stop you even now from committing a crime. But domestic laws and policing mitigate it. Buying a robot and equipping it as you suggest leaves a big data trail. You don't need "a good programmer" to accomplish what you suggest. The barrier isn't technical, it's law and liability driven. Also, fully automatic weapons are not so easy to procure as you suggest. At several steps in your scenario a person is looking at years if not decades of Federal prison time.
As a US citizen I agree, history shows it benefits no population to have concentrated power. That's also why open source is so important in this upcoming era.
It will be just so wonderful when there is widespread access to groups of jerks who will cook up all kinds of chaos buttressed by the power of AI, right? I'm not arguing that only the US should have the tech, but rather that the AI revolution will end up a horrible mess for humanity. Yeah, I'm a doomer.
If im not mistaken, in this same podcast, Schmidt also mentions that China trained Deepseek with like 3M dollars compared to us in the US using Billions of dollars to train our models. Our chip embargo on them has backfired. They have found out how to make improvements and do way more with way less.
@eSKAone- Exactly. Actually, now that I'm thinking about it, I think it may have been a recent podcast I saw with Kai Fu Lee and not Schmidt. Either way, China is finding ways to make engineering and architectural breakthroughs and efficiencies and not relying on pure scaling
@@wedding_photography ua-cam.com/video/Xt4cMYg43cA/v-deo.htmlsi=zJGSjIboqv5Ozc1w It was Kai Fu Lee at this conference. Skip to 20:00. But my mistake, it's his own company's model and not Deepseek.
Yup, they warned Trump not to do it. You want people dependent on you. Unfortunately, when you act without thinking, you mess up. Now we're dependent on them for tons of things and the market. I just hope Trump doesn't make more rash decisions like this when he takes the reigns.
The saftey of AI will come down to the maturity of humanity. And I'm afraid we are not ready. Not even close. We haven't even figured out how to share land and resources without conflict.
We are not being remotely safe. If we actually get a superintelligence, we have to hope it doesn't have a failure of friendliness. Because about 2 minutes after it emerges, it's going to be much smarter than humans. We'll be like dogs or even flatworms to it. It'll be like a 2 year old playing with chess pieces against a grandmaster at chess within the first 30 seconds.
@@DIYSEC The terms, "Friendliness" and "Failure of Friendliness" are terms of art. Failure of friendliness in the context of artificial intelligence (AI) refers to a situation where an AI system does not act in ways that are safe, ethical, or aligned with human values. It's a concern that arises when an AI, despite being programmed with good intentions or goals, misinterprets or misapplies those instructions, leading to harmful outcomes.
@@DIYSEC I recommend reading the book "Superintelligence: Paths, Dangers, Strategies" was written by Nick Bostrom. Published in 2014, the book explores the potential future developments of artificial intelligence and the implications of creating machines that surpass human intelligence. Bostrom discusses the challenges and risks associated with the emergence of superintelligent AI and emphasizes the importance of ensuring that such technology aligns with human values and goals.
Good to hear someone was genuinely SHOCKED in the real, and not just in video title! But can this man be trusted, and maybe he was only surprised, or mildly taken a-back?
Being outside the US, people are speaking up NOT to trust a convicted felon as a leader. There WILL BE NO treaty!!! You can dream but it will not happen.
Peacetime treaties and agreements are trashed all of the time. Think about DPRK and Iran and nukes. Look at CCP and corona virus research and UN judicial ruling on South China Sea. The U.S. cluster bombs on Ukraine raised a few eyebrows. I'm not picking on anyone, just pointing out the way things have worked out during relatively peaceful times. @@Iightbeing
The masses are losing their ability to think critically. Instead they lean on a kind of social media mob mentality for “truth” and “answers” that then self replicate like a lie virus.
Is it? Haven't we seen the same with crypto influencers since 2018 atleast? Scams? The only danger there will be very young people and inexperienced people in tech getting taken by it once but then they will learn instantly how to spot it.
China have some of the most intelligent people in the world and the ‘ban’ on Nvidia gpu’s hasn’t stopped China using them for training. Think about what China and the US aren’t releasing!
The 10,000-person-problem ( 13:50 ) ... is precisely what 3-letter agencies and big-tech themselves are doing, right now. BUT, they want to prevent you and me from having the same ability.
If you use a youtube video in your video, would you mind linking the video in the description so we can check out the entire video after yours? I always end up trying to find the source videos and sometimes they aren't easy to find :(
I think the Westerners have a very narrow understanding of the field of AI, the majority of the teams currently working on AI are Asian mainly Chinese or Indians, why would anyone be surprised that they are releasing the best models ? If you follow the scientific papers released on AI most of them comes from Chinese teams and even in OpenAi a huge percentage of the team is Chinese. I think that China is truly the leader in terms of research on AI even if the most mainstream products are from the US.
@@i_accept_all_cookies He's just looking to get in on the bonanza around conflict with China. All these creeps, finks, and spooks are snarfling at the trough.
Well, only a handful of intelligent psychopaths on top, will have full access. You and I get heavily monitored, sandboxed and rubberized AI ... just smart enough to hook you, but programmed to always stay inside the box. When owners of Big AI (and government) talk "AI safety", they actually mean keeping them safe from the rest of us ... as in: _AI must never help the riffraff escape control._
openai were not the first to use test time compute, anthropic were. remember the token? that only got out once someone figured out how to display it? they didnt invent it either. chain of thought was known already.
Chain of thought was also a lot of hype (like a lot of OpenAI things nowadays). It doesn't improve 90% of the problems. It only excels at 10% of the problems compared to other models, but uses 10x the tokens and thus is costlier. Nobody tells you but all AI companies are losing money. OpenAI would have been bankrupted within a year if it wasn't saved by Amazon, Microsoft and Apple's investment in 2024. Losing $700k a day (the estimate must have gone up by now to $1 mil to $1.5 mil) just for inference and training (not including other operational costs like employee salaries).
The United Nations has been discussing the required involvement of Human-Machine Interaction (HMI) for management of Lethal Autonomous Weapon Systems (LAWS) since 2019 - they have targeted ratification by 2026 (ref UN "2023 New Agenda for Peace")
Hi, that was a great AI news video, I have a question, what is the most interesting thing you have learn in these 3 AI books "Genesis etc..." ? I want to know this because it will be really interesting for me to get your opinion.
just looking at this video and all the conversation in this comments just make me sick. it is non-stop toxic discussion of who is better than who and non-stop China vs USA toxic discussion.
Every time there is even the tiniest advancement in LLM's I get excited! I want to have a lil robot friend that doesn't feel half baked, and every day we seem closer to that. Exciting stuff!
@@Rollthered I'm waiting for a robo-slave I can get to do all my weekend DIY projects with 😂 Imagine having your own team of labourers that will do all the grunt work for free and never complain about needing longer break times 😂 You could build that new garage or shed in no-time
This whole controversy about China capacities in IA is the proof of a real complacency on the part of certain Americans. Chinese researchers have been involved in the domain for more than 10 years. You have Chinese researchers co-authoring most of the major breakthrough in IA. Many of the important papers in the domain have Chinese authors. On the other hand, there are very _few Americans_ that have had a major impact on the domain. Most of the people working at Open AI are either Asian or European. The Americans are nowhere to be seen. This is due to the structure of the education system. Going into math for instance doesn't pay the bills and the brightest Americans usually study to become lawyers. The US rely heavily on the external world to fill in these positions in computer science, mathematics and hard science in general. When IA was a prospect, nobody in the US thought that having Chinese students involved at every stage of this research was a problem. Now, this is a different story. But let's be clear: _The Chinese have been here from the beginning;_
And even the ones who do go into math often end up as quants, helping rich people become even richer by manipulating the stock market, rather than contribute something of value to the world.
Why do you miss the fact that many people of Chinese origin are Americans? China is bleeding people right now. Thousands of them come to the United States.
The thing is that even if they got Ai to cooperate with us, there's no telling how incredibly intelligent it will get. Point being that on the short term of 100 years, it could tell the smallest of lies somewhere that will ultimately work out in its favor over the next 500 - 1000 years. Something that we would never detect. It doesn't need to work on human lifetime scales, other than for it to give us enough reason to not pull the plug on it.
Honestly never really thought about that point of view. Very valid. But that's assuming it had malicious intent and held that and had a long term plan. That's a lot of context to fit in a model lol.
What are these assumptions that China can only compete through reverse engineering or stealing ? It used to be the case 10 - 20 years ago, but now they're maybe the most productive country in terms of research.
Yeah its just coincidence these sort of things pop up right after releases in the west over and over. I mean in the west you see iritative progress over years in china in just pops up after companies in the west has devloped something. Same with robotics. Sure china researches but it's chepaer to let someone else do R&D and then get a hold of that R&d. Can't blame them. It is probably the way to go for the west in the other direction too since its too difficult to keep things hidden. Might speed up progress for everyone anyway
Just take a look at their automotive industry. It's kinda rife with good examples. Check out the LandWind7 SUV. Bruh. I think it's a cultural difference, they just don't have as much issue with aspect of "That's obviously ripped off" if you can get it cheaper as well.
@@gubzsSo the hostile narcissist is afraid that those of his ilk will be superseded by a more advanced intelligence that has the possibility to level the playing field for everyone, thereby sidelining their hostility and exploitative behaviour? Sounds about right 😂 AI could be the holy grail of easy street for everyone, OR it could usher in the authoritarian techno dystopia we see in many movies 😂
@@scroopynooperz9051 Definitively the latter. You can be 100% certain, that current elite are wielding everything they got now, to get there first, and foolproof their position on top, forever.
Wonder how much of this “breakthrough” was directly related to some level of corporate espionage? Many of our top technologies seem strikingly similar to what we are building in China and then they have a “discovery” and amazingly it works and looks & works almost exact the same.
I would wholeheartedly welcome our AI overlords, as they would bring a refreshing change to leadership. I'm exhausted with the current state of human governance and the plethora of rules that come with it.
@@2ndbestfriend Meanwhile DARPA is developing autonomous killing robotic systems, so you're good with that too, eh? So you are just assuming ultimately that ASI will be benevolent? What are you basing that whopping assumption on?
Experts who have "seen some more stuff" than the rest of us are not necessarily going to have more accurate predictions or deeper insights in a nascent field, but they might, so it is worth considering what they have to say.
We'd usually be reverse engineering every step of the way, just like everything else... especially the main thing, the *_"field"_* that influences, provokes, entrains and controls the bot _experience_ ... So just what exactly is the goal again?
Don't believe all the hype. We don't have what we need for AGI: A lot of research in quantum AI. Google is doing that but we are decades away from AGI. The current transformer based LLMs are not progressing linearly as before in intelligence. And we have exhausted all the data online and given to these models.
The beacon of moral light Eric Schmidt tells you what is good for you because he was executive chairman at Google/ Alphabet Inc to 2020 who was in charge when they first changed the company motto from ‘Don’t Be Evil’ to “do the right thing” and removed ‘Don’t Be Evil’ in 2018 at last. Yeah best guy say what is wrong with the world and the best solution giver to "fix" things.
Thank you. People like Wes are not real journalists and uncritically accept self-serving quotes from the likes of Eric Schmidt-one of many people (such as Musk, Altman, and many more) whose pronouncements should always be greeted skeptically.
@@JohnSmith762A11B Sure. I don't know if Wes thinks of himself as a journalist. I take what he does more in the job description as a reporter. Sadly he doesn't put what he reports on into context, so yeah. But on the high side, not having much "put into context" is at least more helpful than having opinion pieces by "journalists" who do not declare it is opinion / bias or interests and the crap they put what is watermark for todays low effort mainstream journalism.
I was shocked to hear Eric's interview and discover his criminal and terrorist mindset. If he had talked with his employees (instead of talking down to them), he would have certainly discovered that most of them were not born in the U.S. You can also notice this by opening a random AI paper and reading the names. So I'm pretty sure he was shocked to discover that people in other countries have intelligence too. But I guess this is hard to grasp for arms dealers and criminals who want to destroy and kill people in every other country with the help of AI drones, yet see open-source models that benefit humanity as an existential threat...
this year our comany had a meeting with a chinese organisation that offered us a box that would run 500 personalities that each have all the social media accounts
Who is working on the authentication problem? I know companies don't care as long as advertisers spend advertising to bots. But users will come around eventually and people will want to communicate with people.
Using open-source software is not stealing, and backwards engineering is legit. Don't make baseless attacks unless you have proof of something illegal going on.
I think social media is like a bucket with holes in it, and AI is water poured into the bucket. Suddenly, the leaks are a lot more apparent, and a bigger deal. I believe this will force social media companies to take bots seriously, because if there's two networks where one has mostly humans and the other has a 10,000:1 bot-to-human ratio, which one are you going to want to use? (I mean bots pretending to be humans. I don't mind bots that admit they are bots.)
It might force government backed identification (like in South Korea or China). That will exclude a portion of the population and create two circles of social media. The ones with gov backed ids, without too much spam and limited foreign interference. The one with all the weird shit. The honest question is, which one of those two circles will have the most influence?
This would be impressive if those deep seek numbers weren't mostly fabricated. Independent testing doesn't support that level of competence or accuracy.
There goes the US thinking everyone is behind until Russia tests a Tsar bomber at 50% capability, but 50 times the potential destruction. Then it's treaty this, embargo that, after they'd killed thousands and affected millions, in Hiroshima and Nagasaki. Flippin hypocrites.
@15:02 Well, hold the phone. You know California will find a way to slow the AI agents down to a grinding halt so that it takes the AI apartment housing agents a very long time to accomplish anything.
A treaty to prevent automated weapons w/o an operator to greenlight it is naive. We have to remember that AI will be smarter, and more aware than the average human. So it doesn't matter if we agree on something, the AIs will have a seat at the table, and they'll have their own agendas and self interests to look after.
Unfortunately with so many people giving up their personal info and photos/videos. Starting with parents posting their kids from the start. I see a grim future for the newer generations having their power of choice. They will easily be manipulated in believing its their choosing by those that wield the coming power. You can argue its been happening already. But just sayin its looking worse.
It will become increasingly more difficult for humans, when they take a look under the hood at any given time slice - to detect deliberate obfuscation by the AI system in question. Even now with human assisted learning - it's often impossible to determine how the system arrived at its final conclusion, solution or answer. As system complexity, interoperability and random communication pathways exponentially ramp up - humans will inevitably find themselves increasingly pushed to the sidelines - even with complete oversight and the best intentions. Mistakes will therefore be made - the size of these mistakes will grow expontially as the complexity and power of these systems increases. How then can anyone be held accountable?! "It wasn't our fault. We did our due diligence but failed to see or comprehend any sub-version and the hidden goal until it was too late".
DeepSeek did what OpenAi supposed to do!
And Qwen
It’s morbidly hilarious that they think this genie is going anywhere near that bottle again. This is totally unstoppable now.
i dream of genie lol.
I get what you're trying to say that at this point we can't put AI back into the Genie bottle which is so true. What I say is screw it it is a very useful tool now we just have to figure out how to harness it and try to stop it from being used by Bad actors and do what we can
@@Timely-ud4rm thats easy make ai a follower of Jezus and abide to Gods laws not a house build om loose sand! Problem is many AI devs are atheists or worse ;)
@@mpwebwerk What are you on about? AI is a tool to be used and harnessed. Thats like saying you want your toaster to follow the laws of god it's ridulious. Keep spiritual stuff humans and humans only and leave AI out of it what's next your gonna try to convince your shovel that it should believe in god? I am spiritual myself but AI isn't even alive or conscious for that matter it's just a sorting algorisiom that been trained on data other than what videos to recommend and what is the genome.
@@mpwebwerk This is why we need AI... to bitch slap some sense into people with their backward medieval BS ideas.
As a non US citizen it would concern me more if only the US had capable models. We need all the countries in world to be participating so we strike a balance. Concentrated power is not good.
A robot can be had for a few thousand dollars. An automatic weapon for a few thousand more. The weapon needs no software. The robot is a networked device with edge computing capability. A good programmer can create the lethal combination. Treaties will do little to stop this from happening.
Lol that's what worries the hegemonic status quo rulers - the fact that they aren't the ones with outright tech superiority anymore.
Notice the whole "China is a problem" mantra only started being parroted in their controlled MSM when the Chingiez figured out how to play the economics & tech game and started competing there.
No one ever complained about "over capacity" before as that was left up to the market to correct lol
@@frederickwelsh treaties don't stop you even now from committing a crime. But domestic laws and policing mitigate it. Buying a robot and equipping it as you suggest leaves a big data trail. You don't need "a good programmer" to accomplish what you suggest. The barrier isn't technical, it's law and liability driven. Also, fully automatic weapons are not so easy to procure as you suggest. At several steps in your scenario a person is looking at years if not decades of Federal prison time.
As a US citizen I agree, history shows it benefits no population to have concentrated power. That's also why open source is so important in this upcoming era.
It will be just so wonderful when there is widespread access to groups of jerks who will cook up all kinds of chaos buttressed by the power of AI, right? I'm not arguing that only the US should have the tech, but rather that the AI revolution will end up a horrible mess for humanity. Yeah, I'm a doomer.
If im not mistaken, in this same podcast, Schmidt also mentions that China trained Deepseek with like 3M dollars compared to us in the US using Billions of dollars to train our models. Our chip embargo on them has backfired. They have found out how to make improvements and do way more with way less.
Basically the PS5 Pro problem. If you give developers better hardware sooner they get less creative.
@eSKAone- Exactly. Actually, now that I'm thinking about it, I think it may have been a recent podcast I saw with Kai Fu Lee and not Schmidt. Either way, China is finding ways to make engineering and architectural breakthroughs and efficiencies and not relying on pure scaling
That's probably disinfo. If they could train such a good model for just $3M, why not spend $300M and absolutely dominate the AI world?
@@wedding_photography ua-cam.com/video/Xt4cMYg43cA/v-deo.htmlsi=zJGSjIboqv5Ozc1w It was Kai Fu Lee at this conference. Skip to 20:00. But my mistake, it's his own company's model and not Deepseek.
Yup, they warned Trump not to do it. You want people dependent on you. Unfortunately, when you act without thinking, you mess up. Now we're dependent on them for tons of things and the market. I just hope Trump doesn't make more rash decisions like this when he takes the reigns.
FYI, Kissinger was a psychopath who was, at least, partly responsible for some of the worst massacres of the late 20th century.
Very much so. He also was part of the WEF. Phycopaths trying to take over thecworld. If they get AI we all are done
There is no good guy in statecraft.
Lol and that same Kissinger won a Nobel prize for peace - food for thought 😂
@@scroopynooperz9051 True. So did Obama.
What does this comment have to do with weaponized robots?
The saftey of AI will come down to the maturity of humanity. And I'm afraid we are not ready. Not even close. We haven't even figured out how to share land and resources without conflict.
💯
We are not being remotely safe. If we actually get a superintelligence, we have to hope it doesn't have a failure of friendliness. Because about 2 minutes after it emerges, it's going to be much smarter than humans. We'll be like dogs or even flatworms to it. It'll be like a 2 year old playing with chess pieces against a grandmaster at chess within the first 30 seconds.
@@macmcleod1188 I don’t think at the super intelligent level it will think in terms of “friendly” Useful or useless more likely lol
@@DIYSEC The terms, "Friendliness" and "Failure of Friendliness" are terms of art.
Failure of friendliness in the context of artificial intelligence (AI) refers to a situation where an AI system does not act in ways that are safe, ethical, or aligned with human values. It's a concern that arises when an AI, despite being programmed with good intentions or goals, misinterprets or misapplies those instructions, leading to harmful outcomes.
@@DIYSEC I recommend reading the book "Superintelligence: Paths, Dangers, Strategies" was written by Nick Bostrom.
Published in 2014, the book explores the potential future developments of artificial intelligence and the implications of creating machines that surpass human intelligence.
Bostrom discusses the challenges and risks associated with the emergence of superintelligent AI and emphasizes the importance of ensuring that such technology aligns with human values and goals.
aitranslations AI fixes this. Ex-Google CEO shocked by AI
Him being shocked doesn't surprise me. He's one of the "in 100 years this will happen" followed be "oh wow it'll actually happen in 50 years" guys
That hardly sums up this jerk who has his tentacles all over this tech and already making tons of cash investing in military applications.
12:23 Dead internet theory is a prophecy, not a conspiracy theory.
Good to hear someone was genuinely SHOCKED in the real, and not just in video title!
But can this man be trusted, and maybe he was only surprised, or mildly taken a-back?
Can this man, best friend of Kissinger be trusted? No.
FINALLY, a _PROPER_ use-case for the word "SHOCKED" in the title. THANK YOU!
There is no way on earth any country with the current geopolitical situation will ever trust an adversary and follow a proposed treaty.
Exactly. There will be no rules in full scale war.
Being outside the US, people are speaking up NOT to trust a convicted felon as a leader. There WILL BE NO treaty!!! You can dream but it will not happen.
I disagree.
Peacetime treaties and agreements are trashed all of the time. Think about DPRK and Iran and nukes. Look at CCP and corona virus research and UN judicial ruling on South China Sea. The U.S. cluster bombs on Ukraine raised a few eyebrows. I'm not picking on anyone, just pointing out the way things have worked out during relatively peaceful times. @@Iightbeing
@@eSKAone-I think even during "peace time" some countries will use treaties and agreements to buy time or gain advantage.
11:30 How about we agree to stop killing each other as humans too?
you find jews are receptible to such reason?
That would never happen
Only once we get rid of money. Make everything free. RBE
@@There-Is-No-Virusthen you have forced labor to make everything. Free bare essentials and housing maybe.
@@There-Is-No-Virus you first: give us your time, money, resources, labor, mind for free. oh wait it's always "someone else" isn't it
What he said about the ability to create 10,000 influencers is chilling.
The masses are losing their ability to think critically. Instead they lean on a kind of social media mob mentality for “truth” and “answers” that then self replicate like a lie virus.
it's all BS, if he can, SHOW IT, don't give useless papers.
Is it? Haven't we seen the same with crypto influencers since 2018 atleast? Scams? The only danger there will be very young people and inexperienced people in tech getting taken by it once but then they will learn instantly how to spot it.
@@hqcart1seen it in action myself. see my other comment
@hqcart1 l0l0l0l you've never worked with agents I see.
So shocking that even Eric is shocked? That's shocking!
China have some of the most intelligent people in the world and the ‘ban’ on Nvidia gpu’s hasn’t stopped China using them for training.
Think about what China and the US aren’t releasing!
I'm glad he clarified his point of view. His last talk my by butt pucker when he said "offensive AI."
"Would you like to play a game?"
"The only winning move is not to play."
Too late.
Enjoy being alive today.
The 10,000-person-problem ( 13:50 )
... is precisely what 3-letter agencies and big-tech themselves are doing, right now.
BUT, they want to prevent you and me from having the same ability.
If you use a youtube video in your video, would you mind linking the video in the description so we can check out the entire video after yours? I always end up trying to find the source videos and sometimes they aren't easy to find :(
I think the Westerners have a very narrow understanding of the field of AI, the majority of the teams currently working on AI are Asian mainly Chinese or Indians, why would anyone be surprised that they are releasing the best models ?
If you follow the scientific papers released on AI most of them comes from Chinese teams and even in OpenAi a huge percentage of the team is Chinese.
I think that China is truly the leader in terms of research on AI even if the most mainstream products are from the US.
No, because the best researchers and the leaders of these companies are Westerners.
@@johnalf-yd2uw 😂 we're not paying attention are we
You're probably right, and that Schmidt would say China is one year behind seems willfully ignorant of the facts as they stand.
@@i_accept_all_cookies He's just looking to get in on the bonanza around conflict with China. All these creeps, finks, and spooks are snarfling at the trough.
Kissinger was an idiot. Like and subscribe if you agree.
Shocking what happens if everyone gets hands on this powerful systems and only what matters is creativity, big corporations are in horror.
Well, only a handful of intelligent psychopaths on top, will have full access.
You and I get heavily monitored, sandboxed and rubberized AI ... just smart enough to hook you, but programmed to always stay inside the box.
When owners of Big AI (and government) talk "AI safety", they actually mean keeping them safe from the rest of us ... as in: _AI must never help the riffraff escape control._
openai were not the first to use test time compute, anthropic were. remember the token? that only got out once someone figured out how to display it? they didnt invent it either. chain of thought was known already.
Chain of thought was also a lot of hype (like a lot of OpenAI things nowadays). It doesn't improve 90% of the problems. It only excels at 10% of the problems compared to other models, but uses 10x the tokens and thus is costlier. Nobody tells you but all AI companies are losing money. OpenAI would have been bankrupted within a year if it wasn't saved by Amazon, Microsoft and Apple's investment in 2024. Losing $700k a day (the estimate must have gone up by now to $1 mil to $1.5 mil) just for inference and training (not including other operational costs like employee salaries).
@@SahilP2648Amazon lost money for decades.
That’s Sir Mark Rylance, a brilliant British actor, most might know him from Bridge of Spies or Ready Player One.
Agents is phenomenal.
classic WR banger, TY
video starts at 14:40
This is actually shocking.😱
Charlie Brooker is taking notes furiously
What is the game or simulation shown at 18:15?
The Italian audio track is missing.
NONE of this should be an IP race. This needs a world treaty with all players and non-players at the table. This is where capitalism fails.
The United Nations has been discussing the required involvement of Human-Machine Interaction (HMI) for management of Lethal Autonomous Weapon Systems (LAWS) since 2019 - they have targeted ratification by 2026 (ref UN "2023 New Agenda for Peace")
Hi, that was a great AI news video, I have a question, what is the most interesting thing you have learn in these 3 AI books "Genesis etc..." ?
I want to know this because it will be really interesting for me to get your opinion.
just looking at this video and all the conversation in this comments just make me sick. it is non-stop toxic discussion of who is better than who and non-stop China vs USA toxic discussion.
Now hearing all this news. How will the future be ??? It sound truly scary.
Every time there is even the tiniest advancement in LLM's I get excited!
I want to have a lil robot friend that doesn't feel half baked, and every day we seem closer to that. Exciting stuff!
I'm waiting for a robo-girl friend. No worries about alimony, cheating, stds, etc.
@@Rollthered I'm waiting for a robo-slave I can get to do all my weekend DIY projects with 😂
Imagine having your own team of labourers that will do all the grunt work for free and never complain about needing longer break times 😂
You could build that new garage or shed in no-time
@scroopynooperz9051 damn when they go rouge your first on the chop n block 🫡
Please cover socratic deepmind learning
What's more concerning is the work Whitney Web discussed a year ago about Schmidt with regard to the book he wrote with Kissinger
This whole controversy about China capacities in IA is the proof of a real complacency on the part of certain Americans. Chinese researchers have been involved in the domain for more than 10 years. You have Chinese researchers co-authoring most of the major breakthrough in IA. Many of the important papers in the domain have Chinese authors. On the other hand, there are very _few Americans_ that have had a major impact on the domain. Most of the people working at Open AI are either Asian or European. The Americans are nowhere to be seen. This is due to the structure of the education system. Going into math for instance doesn't pay the bills and the brightest Americans usually study to become lawyers. The US rely heavily on the external world to fill in these positions in computer science, mathematics and hard science in general. When IA was a prospect, nobody in the US thought that having Chinese students involved at every stage of this research was a problem. Now, this is a different story. But let's be clear: _The Chinese have been here from the beginning;_
Absolutely, if you count every notable AI researchers who have a influential paper by its ethnicity, Chinese would be the overwhelming majority...
And even the ones who do go into math often end up as quants, helping rich people become even richer by manipulating the stock market, rather than contribute something of value to the world.
what?
is this a troll comment?
the current AI boom was kicked off by USA lead research papers
Why do you miss the fact that many people of Chinese origin are Americans? China is bleeding people right now. Thousands of them come to the United States.
The thing is that even if they got Ai to cooperate with us, there's no telling how incredibly intelligent it will get. Point being that on the short term of 100 years, it could tell the smallest of lies somewhere that will ultimately work out in its favor over the next 500 - 1000 years. Something that we would never detect. It doesn't need to work on human lifetime scales, other than for it to give us enough reason to not pull the plug on it.
Honestly never really thought about that point of view. Very valid. But that's assuming it had malicious intent and held that and had a long term plan. That's a lot of context to fit in a model lol.
@Isaac-mh4yt basically Ai using plausible deniability
What are these assumptions that China can only compete through reverse engineering or stealing ? It used to be the case 10 - 20 years ago, but now they're maybe the most productive country in terms of research.
Yeah its just coincidence these sort of things pop up right after releases in the west over and over. I mean in the west you see iritative progress over years in china in just pops up after companies in the west has devloped something. Same with robotics. Sure china researches but it's chepaer to let someone else do R&D and then get a hold of that R&d. Can't blame them. It is probably the way to go for the west in the other direction too since its too difficult to keep things hidden. Might speed up progress for everyone anyway
Just take a look at their automotive industry. It's kinda rife with good examples.
Check out the LandWind7 SUV. Bruh.
I think it's a cultural difference, they just don't have as much issue with aspect of "That's obviously ripped off" if you can get it cheaper as well.
I tried deep seek the other night and I can confirm that it is just as good as 01 preview
Eric Schmidt is one of those people that you can just watch while he speaks and tell there's something deeply hostile and narcissistic underneath.
What has this anything to do with the content of the video? In this video he is the one who warns against hostile uses.
@@RolandPihlakas The character of the person delivering the message contextualizes their reasons for delivering it
@@gubzsSo the hostile narcissist is afraid that those of his ilk will be superseded by a more advanced intelligence that has the possibility to level the playing field for everyone, thereby sidelining their hostility and exploitative behaviour?
Sounds about right 😂
AI could be the holy grail of easy street for everyone, OR it could usher in the authoritarian techno dystopia we see in many movies 😂
@@scroopynooperz9051 Definitively the latter.
You can be 100% certain, that current elite are wielding everything they got now, to get there first, and foolproof their position on top, forever.
You are projecting, sir
Wonder how much of this “breakthrough” was directly related to some level of corporate espionage? Many of our top technologies seem strikingly similar to what we are building in China and then they have a “discovery” and amazingly it works and looks & works almost exact the same.
Chinese ingenuity is the ability to get hold of cutting edge western technology by being its prime manufacture on the cheap.
I would wholeheartedly welcome our AI overlords, as they would bring a refreshing change to leadership. I'm exhausted with the current state of human governance and the plethora of rules that come with it.
@@AzogticMettroskik I’m in this camp as well. Powerful AI in the hands of people is scary. But autonomous AI taking over control - I’d invite it.
@@2ndbestfriend Meanwhile DARPA is developing autonomous killing robotic systems, so you're good with that too, eh? So you are just assuming ultimately that ASI will be benevolent? What are you basing that whopping assumption on?
The secret sauce is common sense.
More specifically, critical thinking.
It's literally the steps you learned in grade school.
9:59 have you watch iRobot? Elysium? Or the latest terminator? There are many movies where robot are not slow at all
Experts who have "seen some more stuff" than the rest of us are not necessarily going to have more accurate predictions or deeper insights in a nascent field, but they might, so it is worth considering what they have to say.
That 32B QwQ is unreal.
Care to elaborate?
If you can run it try it out. It's responses are amazing.
it suck, i tried it.
The above two comments directly contradict each other and is a perfect example of why I cannot trust humans 😂😂
@@WhatIsRealAnymoreno it genuinely sucks, at its best it still sucks more than gpt 3.5 on a bad day
Our worst fears about Robots will become our reality, Fear is the only way to correct humanity.
Schmidt is saying we need a treaty against automated systems -- funny thing is we ALREADY have that.
We have what? The treaty or the automated systems?
No one has a MOAT!
HE SAID IT HE SAID IT
We'd usually be reverse engineering every step of the way, just like everything else... especially the main thing, the *_"field"_* that influences, provokes, entrains and controls the bot _experience_ ... So just what exactly is the goal again?
China also owns Hailuo Minimax and Kling AI which are the best video generator tools imo
Did an AI agent also design those 3 high chairs? I think it got the scaling comically wrong lol 😂
That amphibious robot reminds me a lot of updog.
What's updog?
@@AutomatedLiving09not much how bout u?
Leon aka Linwei Ding = source ?
So is this how Star Wars started? - autonomous stormtroopers changed the game
They want to get to AGI as fast as possible, so the more people that use there model, the faster they will learn how to create AGI, right?
Don't believe all the hype. We don't have what we need for AGI: A lot of research in quantum AI. Google is doing that but we are decades away from AGI. The current transformer based LLMs are not progressing linearly as before in intelligence. And we have exhausted all the data online and given to these models.
@@SahilP2648
The beacon of moral light Eric Schmidt tells you what is good for you because he was executive chairman at Google/ Alphabet Inc to 2020 who was in charge when they first changed the company motto from ‘Don’t Be Evil’ to “do the right thing” and removed ‘Don’t Be Evil’ in 2018 at last. Yeah best guy say what is wrong with the world and the best solution giver to "fix" things.
Thank you. People like Wes are not real journalists and uncritically accept self-serving quotes from the likes of Eric Schmidt-one of many people (such as Musk, Altman, and many more) whose pronouncements should always be greeted skeptically.
@@JohnSmith762A11B Sure. I don't know if Wes thinks of himself as a journalist. I take what he does more in the job description as a reporter. Sadly he doesn't put what he reports on into context, so yeah. But on the high side, not having much "put into context" is at least more helpful than having opinion pieces by "journalists" who do not declare it is opinion / bias or interests and the crap they put what is watermark for todays low effort mainstream journalism.
The argument stands on it own merit. Everything else is a distraction.
Wow, that Flux girl...
There is no possibility of a treaty stopping these robots from being created. They are too cheap. The technology is too wide spread.
Someone read Asimov and did not understand that it was fiction.
I was shocked to hear Eric's interview and discover his criminal and terrorist mindset. If he had talked with his employees (instead of talking down to them), he would have certainly discovered that most of them were not born in the U.S. You can also notice this by opening a random AI paper and reading the names. So I'm pretty sure he was shocked to discover that people in other countries have intelligence too. But I guess this is hard to grasp for arms dealers and criminals who want to destroy and kill people in every other country with the help of AI drones, yet see open-source models that benefit humanity as an existential threat...
Exactly 👍 😅
enjoy going outside while you can. as long as the military operates like this, its only a matter of time before public space becomes dangerous.
this year our comany had a meeting with a chinese organisation that offered us a box that would run 500 personalities that each have all the social media accounts
the reason it was a box is because you also need to spoof 500 device profiles to get past twitter etc. bot protection
box fits in a standard server slot in a rack and costs 40k Rmb. my company ordered 50 of them
i welcome ai military robots, however i believe the treaty should state that no one should send human soldiers to fight against robots
Oh cool so we don’t need ID to sign up for social media ? We can just 1 click 10k
at least they made it open source , who do you trust OpenAI or China?
None of them, but CCP is evil
Open AI
OpenAI
After the Industry was shocked multiple Times, finally someone within the industry admitts he ist shocked!
Schmidt is only pretending to be shocked to gain some regulatory advantage for one or more of the companies he is invested in.
Why the audio is in French?
Go to Settings > Audio Track and you can select among several (AI translated) languages
Thank you Wes. This channel is always my go-to for all things Ai.
Who is working on the authentication problem? I know companies don't care as long as advertisers spend advertising to bots. But users will come around eventually and people will want to communicate with people.
11:37 Terrible examples. Even with an agreement in place, there's no way to tell if it was Ai or people.
Always stellar presentations Wes, thank you
"Used to work for Google"... was CEO of Google :P
CEO in title ... COO in reality.
@@coldlyanalytical1351 No big mouth ceos anywhere
Using open-source software is not stealing, and backwards engineering is legit. Don't make baseless attacks unless you have proof of something illegal going on.
Did I just hear that "our" computer scientists just "cracked" an AI model whose code is open source?😂
How can we trust our own people, much less others, to carry out agreements & regulations in good faith?
I think social media is like a bucket with holes in it, and AI is water poured into the bucket. Suddenly, the leaks are a lot more apparent, and a bigger deal. I believe this will force social media companies to take bots seriously, because if there's two networks where one has mostly humans and the other has a 10,000:1 bot-to-human ratio, which one are you going to want to use? (I mean bots pretending to be humans. I don't mind bots that admit they are bots.)
It might force government backed identification (like in South Korea or China). That will exclude a portion of the population and create two circles of social media. The ones with gov backed ids, without too much spam and limited foreign interference. The one with all the weird shit. The honest question is, which one of those two circles will have the most influence?
This would be impressive if those deep seek numbers weren't mostly fabricated. Independent testing doesn't support that level of competence or accuracy.
Their Arena scores are down near Mistral for God's sake
Most of the new Chinese models are probably heavily fine-tuned llama versions with some rudimentary Chain of Thought slapped on.
There goes the US thinking everyone is behind until Russia tests a Tsar bomber at 50% capability, but 50 times the potential destruction. Then it's treaty this, embargo that, after they'd killed thousands and affected millions, in Hiroshima and Nagasaki. Flippin hypocrites.
If you used deep seek you know how horrible it is
Deepseek R1 solves a simple 10 character ROT17 cypher, can Llama do it?
I’ve given an executive orders to AI that will be followed. You have all been warned over and over again.
@15:02 Well, hold the phone. You know California will find a way to slow the AI agents down to a grinding halt so that it takes the AI apartment housing agents a very long time to accomplish anything.
A treaty to prevent automated weapons w/o an operator to greenlight it is naive. We have to remember that AI will be smarter, and more aware than the average human. So it doesn't matter if we agree on something, the AIs will have a seat at the table, and they'll have their own agendas and self interests to look after.
Ex-Google CEO get’s “ELECTROCUTED” by AI!!!!
Thinking models... if Answer = Crazy then regenerate();
What do we think? It’s the end of social internet
Unfortunately with so many people giving up their personal info and photos/videos. Starting with parents posting their kids from the start. I see a grim future for the newer generations having their power of choice. They will easily be manipulated in believing its their choosing by those that wield the coming power. You can argue its been happening already. But just sayin its looking worse.
My only question is.... how tf is henry kissinger STILL alive?
He died late last year.
@@coldlyanalytical1351 O_o 19:15
Meanwhile their economy is crumbling (its sort of tragic sad).
You hear that Mr. Anderson?... That is the sound of inevitability...
in the dune universe it took a butlerian jihad to get rid of the 'thinking machines'
It will become increasingly more difficult for humans, when they take a look under the hood at any given time slice - to detect deliberate obfuscation by the AI system in question.
Even now with human assisted learning - it's often impossible to determine how the system arrived at its final conclusion, solution or answer.
As system complexity, interoperability and random communication pathways exponentially ramp up - humans will inevitably find themselves increasingly
pushed to the sidelines - even with complete oversight and the best intentions.
Mistakes will therefore be made - the size of these mistakes will grow expontially as the complexity and power of these systems increases.
How then can anyone be held accountable?!
"It wasn't our fault. We did our due diligence but failed to see or comprehend any sub-version and the hidden goal until it was too late".
Never trust Schmidt
There's a point where he touches the back of his neck or ear meaning he's either lying or legit scared.
No, that's not what that means.
shocked!