Stunning AI shows how it would kill 90%. w Elon Musk.

Thanks to Brilliant for sponsoring this video.

I used the GPT-3 AI and a Synthesia avatar (plus DeepMind’s flamingo).

For a deeper drive, I'd recommend the book Life 3.0, by MIT Professor Max Tegmark.

Sources

DeepMind CEO Demis Hassabis, with Lex Fridman

Elon Musk, with Joe Rogan

Detecting consciousness, Cambridge University:

Imaging detects signs of consciousness:

Flamingo AI, DeepMind

Dr. Ben Goertzel - Artificial General Intelligence

Quantum gravity sensors:

“Solving these scaling challenges is what will deliver AGI.”
Nando de Freitas, Research Director, DeepMind (former Oxford Prof)

CGP Grey:

[DigitalEngine]

Also, we're hiring. If you're interested, please get in touch via the about page.
Thanks!

[katelynwoodworth]

So let me get this straight... we have NO idea what AI will do to humanity in the future, and yet we are racing forward to make it... has anyone thought about feedback systems where the AI would feel guilt or anguish over immoral actions? Because as I see it, we are making AI to be psychopaths.

[DaystromDataConcepts]

Nothing I hear about how A.I is developing fills me with a warm feeling inside.

[richardlanier]

My father hated the internet and newer technology. He said it will cause problems for mankind, and that it will dumb people down. He died in 2009. He would be appalled at what is happening now. I was not too sure at the time what to make of it, but now I think he had a very valid point.

[majkels]

The terminator movies were not made to show the scientist where they should draw a line or be carefull. They were made to show us, normal people, that scientists will blindly pursue AI development regardless of the consequences.

[klaushubbertz]

10:58: "AI could remove humans in many ways: Kill us directly, destroy food sources, enslave or manipulate us, create powerful weapons, control the climate, manipulate the weather"
This is a perfect description of current real affairs.

[olif]

Quote from jurassic park comes to mind: "Your scientists were so preoccupied with whether they could, they didn't stop to think if they should."

[oddballskull]

One person quotes jurassic park, we ALL must quote jurassic park

[jbaker]

I don't think hunanity has "collectively decided" to take AI as far as it is NOW, but that doesn't seem to have stopped any scientists yet.

[mpjstuff]

Elon in his speech is saying; "If humanity collectively decides." Well, it's not COLLECTIVELY deciding anything. it's a few people and governments who can afford the toys. If everyone could have access to build AI -- I'd feel better about it than the guiding hand of greed and hubris.

[EccleezyAvicii]

These AI dreams are sorta a glamours version of an actual dream I experienced shortly after I started using chatGPT.

There was a fear so profound I experienced, that I have no way of describing it other than it colossally out-ways every conceivable idea I’ve had throughout my life by orders of magnitude.

In the dream it felt as if I arrived suddenly to a distant future were AI had not only run its coarse in displacing humanity, but that it also resulted in something so disastrous, something that I couldn’t quite grasp until I experienced. It wasn’t order, or beautiful, there was no symmetry, no consistency. It was as if the ultimate reward function prevalent in the fabric of reality all around me was: reshape every atom to be uncanny and corrode anything normal by way of glitch. It was horrifying, it is the most vivid memory I have had in my life and to this day I remember the incomprehensible dread that I sat with that morning.

How could this happen, surely I thought the artificial, spurred by a higher order of intelligence would have wrought a more beautiful and profound world. I realized that was a massive miscalculation, an oversight, an assumption. The reality is that chaos plays a much larger game, and the default future of artificing intelligence is not order.

[andrewoneil]

It’s shocking how many people either ignore or laugh confusingly when AI is brought up. Most have no idea that this could be the last invention humans need create.

[LuisChavez-ehvr]

I find it interesting how we keep hearing how dangerous AI is, but we continue to advance it

[Mr.Nichan]

I feel like the main reason that these AI's have these thoughts (being human, maybe eventually deciding humans are unnecessary and wiping them out, that idea being "a scary thought") is that they are trained on our media, and humans have been speculating about these things and writing fiction about them for a long time.

[wildbill]

"The good thing about computers is that they do what you tell them to do. The bad news is that they do what you tell them to do" - Ted Nelson

[dixienormus]

"How many mirrors are there?" "At least 2" 😂😂😂

[nmdakini]

One thing humans have definitely not learned is just because we can do something doesn’t mean we should.

[jake]

Elon musk once said “AI doesn’t have to be evil or hate us to destroy us. It can just have the logical thinking that destroying us is the best course of action for its desired goal. Just as we don’t hate ants when we crush their home while building a highway. It’s just a necessary thing that we don’t even bat an eye too.” That’s the scary part. If we tell an AI to “help global warming” or something like that, we could be signing our own death wish.

[justacoolguy]

Scientist: AI can kill us in the future
Also scientists: Let's continue to upgrade AI

[Dadsezso]

@11:02 What Elon says here is significant. "If humanity collectively decides" The problem is, it isn't a collective decision. Those developing it and have the resources to create it have not asked the rest of us if we want this. More than once I've seen comments from various groups designing AI that tell us they really don't care what the bulk of us think about what they're making.