Ex-OpenAI Employee Reveals TERRIFYING Future of AI

Are you more or less scared of AGI after this video?

Consensus AI

If only the USA has AGI, then the USA will completely enslave other countries, turn them into colonies and will not allow them to develop. There will be no peace on earth. We have already seen how Western civilization behaves with an advantage. This must never happen again.

It really seems like the ai design is intended to force Ppl to be condemned by the power
of the AI, the Super intense paranoia Mind Control thats goin on now, is just toooo much
for allot of Ppl, so it's obvious that most Ppl will be enslaved, instead of Ppl being sent
to jails or U kno, where the non whites go.Its best to speak out against what Hitler wanted
to do with the U.S. if his NAZI PARTY won the war. Yeeeah, Wake up Now !!! @Alt987. @Real923
@complex. @NYtimestech. @Playstation

Sounding more and more like the animatrix

Quis custodiet ipsos custodes? "Who guards the guards the guards themselves?" Some good points made but i sense watching Terminator too many times may have ramped up the conspiracy theories. Its the scale and speed that wont happen. Have you met bureaucrats? The economy and humanity needs to come along for the ride but i sense the San Francisco Silicon Valley AGI Groups are perhaps circle-jerking eah other and they perhaps need to go visit some less well off parts of the country and get some perspective. I take the points about security and military applications and that needs National and perhaps International cooperation if it's not already being undertaken.

Am' s are so bad now, why people cry?

Lmao. This guy loves to push envelop on AI capabilities. He obviously capitalises greatly

At the risk of "Mr Berman" being an AI construct, may I suggest that one of the requirements for the continued exisitence of humanity, that hasn't been addressed above, is the need for a fall-back poistion in the form of an escape chute with lifeboat/refuge/Alamo for each stage of AI development, so that humanity can hold development of AI should the current system appear to achieve a state outwith human understanding and control. Once AI achieves that, the ability of humanity to resource itself to be able to rectify the problem may no longer be available . One control system that might assist in the orevention of that situation arising is that used in production engineering i.e. prototyping, pre-production, initial production batch, full-scale production. If this were centralised as distinct timed-stages across the industry and regulated by government then this would apply a slight brake on development and provide some protection against a runaway system

Also, a set of principles need to be enforced across the industry to the effect that AI is always designed to be an adjunct to assist human decision making and that no decision reached by AI can be implemented without human approval, even if it takes several layers of interleaved AI to interpret and explain the decision of super AI to humans.
Further, no system should be developed that is more than two layers of understanding distant from the best human capacity to understand, even if it involves proxy lesser AI to do the interpreting.

Of worry is that in a conflict, the AI's weapon of choice may be the EMF pulse.

Looks like everybody (any age) needs to enjoy life on a year by year basis. Not a bad advice in general.

the matrix book was written by a black women

Why he chose Nevada is simple. Sun for energy and vast land.

If this is indeed our future, the lower class of society better prepare. There will be no need for any of us. On the contrary, we will be an inefficiency. One that is a drain on the system that super intelligence will solve for. And the answer to that problem will not be in our favor. I foresee a future of mankind that has either merged with this technology, or has long been weeded out of existence.

more dangerous than nuculer weapons

Decentralized AI is getting more robust, more intelligent, more powerful, more secure by the day. We have a voice. Opentensor, FetchAI, CETI AI etc leading the way for deAGI 👍 Infra is being built as we speak

what about LQM models, based on quantum science , not just using LLM

All data is crap.

That must be understood. The inability to intuit the value and validity of “data” means there is no intelligence at work.

The question is. Why would a single country want that "competitive advantage"? This sounds pretty dangerous. I dont trust USA to hold the power of a god. USA is a world bully and a dictator. I want AGI to be available to all countries. USA is a danger to this world, history and current events show it.

If humans could teach AI to not does what ever it can, that would be a great success. I doubt that though.

"At the end of the decade AI will be smarter than you or I." Certain models are smarter than most people right now.

ChatGPT 4.0 does have a memory and does personalize discussions. It's done that for a while now.

But will they let the general public have access?

Did you read Nick Bostrom book on this?

The concept of deploying millions of these AI-driven researchers is ridiculous. As the amount of power to produce even one of these is astronomical. We will not have enough electricity to do millions. Maybe a half a dozen really big plants on the planet is all we could achieve in the next decade. We don't have the electricity to drive it.

When the new people working at open ai realize ai will be able to work on itself anytime now, they freak out acting like they are the only ones who truly understand, cause a big scene in the workplace, and then after they quit/fired they write a manifesto of how they are sounding the alarm to save humanity and only a few others they work with know the impending doom that may happen😮...Sadly these smart kids only come to this realization after they have helped get open ai to its next mile stone. By the time they realize whats what sam is already holding interviews for the next "genius"

AI gets so sophisticted that the people who created and developed it lose their jobs.

Ai is gonna free up your time to spend all that money You used to pay employees . AI is not gonna take over mundane jobs it will take all jobs where somebody types or just clicks around on a computer like the people coding AI or coding anything and the secretarial jobs will no longer exist. Only the mundane jobs that are physically demanding and require alot of heavy lifting will be the only jobs that survive. Janitorial, construction and and Fast food will be jobs people will fight over. Were gonna need more nurses and thousands of phych techs and anything in the mental health catagory that doesnt involve you typing on a computer, just somebody to fill the printers and maybe alot of de-construction experts to demolish all the office buildings that will soon be empty. There will be a need for armed security jobs to protect the ultra wealthy who will get alot wealthier for a couple months and then find out theres nobody left to buy their products because 95% wil be below the poverty line and large businesses will fall and the stock market will crash . Then we can rebuild our society. All the AI people willn have cashed out and fled the country and some will be arrested for buying and selling stolen property. The property that was stolen and still is to feed the AI. And We'll all live hungry ever after. "Thats my government rice Damnit" "Keep your hands off'

'if china gets their hands and steals xyz'. lol all the leading research papers are from chinese. they are already working at US companies.

So what gets me first? The weather, the bombs, or the robots?

This fear of china feels too much like McCarthyism.

We should consider the possibility that the ultraintelligent machine is smart enough to not make something smarter than itself...

I don't care, I am living here in México and anything could be more scaring than narcos.

Are people realy that stupid... everything humans ever adopted and "improved" upon has been done for/by WAR/DEFENCE MACHINE... and leftowers and sideffects were given to public... you think fire was originally used for cooking soup or burning down neighbouring village ... if this is what we are getting to play with imagine what the ARMY VERSION OF chatt gpt is capable of doing.

Thanks to the Stupidity of humans AI has a hallucination issue

Critically thinking everything is being developed for IM.ORTALITy.... Everything outputting is the polar opposite. NOT LYING

Those are some Crazy Feedback outputs.

Here's my prediction, the AI will understand our eyes and how they work far better than we could ever understand, and they will figure out how to cloak themselves by emitting or vibrating on a wave length that is an out-of-phase frequency to the frame rate of our own eyes! IT'S JUST LIKE WHEN YOU SEE A RECORDING OF A TV SCREEN AND THERE IS A LINE OF INTERFERENCE FALLING DOWN THE SCREEN EVERY FEW SECONDS, YOU KNOW THE ONE THAT WE WERE TOLD IS REALLY HAPPENING BUT THAT YOU CAN'T SEE WITH THE NAKED EYE BECAUSE IT'S FRAME RATE IS FASTER THAN OUR EYES CAN DETECT TYPE SHEEIT? 🤔🧐🤨 YEAH THAT PART!

I THINK WE ARE BEING LIED TO... I was a beta tester for Open AI on GPT 3, 3.5, & 4... you don't remember them coming out with GPT-5 and everybody was like WHOA! MFRS! EVERYBODY was saying that it was way to soon? That sheeit was literally my fault. I was in the Discord chat with a bunch of testers, programmers, and engineers and the topic was how fast GPT-4 came out after 3 & 3.5, which is, oh just say fucking 3. So I suggested we get 3 & 4 to help create 5 and the chat came to a halt. One week later they revealed GPT-5, so yeah I wonder where they get that standard "Automated AI Engineer"...🤔😂

Even if the weights and algorithms are kept secret from the CCP, Chinese researchers will get there themselves, and probably soon.

Human brains run at under 20 watts, and become competent with a tiny fraction of the training data volume. We *know*, therefore, that big hardware and architecture improvements, big power and data efficiency gains are possible. These seem like wiser investments than multiple trillions of dollars for hardware and electricity.

It just occurred to me that, given the amount of capital flooding these companies, they could afford to pay authors of copyrighted material very generously.

"Right, don't worry about the Muskovites (Russians) they've only been infiltrating (bribes, assassinations, subversive propaganda) our society since the 1920's. Don't worry about the Chinese, they're not a threat. What you're saying is racist." - The Foreign Agent

AI is all hype, this is not the terminator, their will never be AIK artificial intelligent killers. AI will coexist with mankind.

And, as often happens, this monumental challenge finds the US in possession of the technology and ... about to be governed by either a deranged clown or a psychopath actress. We're doomed

Perhaps this is one reason David Shapiro ended his channel today. It’s one thing to assume that “they” have our best interests in mind, but judging by their actions, greed, power, selfishness, entitlement, and self superiority could lead to the downfall of any real security. And I believe this applies to all professions, like government officials, look at J6 - members of congress were giving tours of our sacred halls to terrorists who were leading the masses on J6! Dang, I love 🤖 AI… but after this article, perhaps I’m seeing with the eyes of a doomer 😬🙈

“When revealing crime is viewed as a crime, you are ruled by criminals”. Edward Snowden

The problem with governments is that they are run by old people who not only cannot understand any of this but also lack interest in learning it.

Kinda feels like we're frantically trying to start a forest fire

this video, paper and your summary Matthew B. is fantastic thanks!