so who won? did the motion pass or fail?

"To cause us trouble, such misaligned superhuman intelligence needs no robotic body, merely an internet connection - this may enable outsmarting financial markets, out-inventing human researchers, out-manipulating human leaders, and developing weapons we cannot even understand. Even if building robots were physically impossible, a super-intelligent and super-wealthy AI could easily pay or manipulate many humans to unwittingly do its bidding." (Max Tegmark)

"I can drive a car, so i can go to the moon." Isn't that exactly what happened? Engineers kept improving the engine until they went to the moon. I don't think AI has been "singled out" as the only existential risk. Nuclear weapons are considered an existential danger, as is climate change, and pathogens also.

Chat GPT is Microsoft FrontPage 2024 - It's not going to go all Skynet - and it will probably create crappy websites just like Front Page Original.

4:14 that where Ultron comes hahaha

If only history weren't full of great men and scientists being completely wrong, I'd be less worried.

@ZONA_ZERO_OFFICIAL

5 ай бұрын

Is time to educate with ancestors facts … They say that On 2020 we find out that white people are bad … ? Considering that white people are really the artificial humans such as LLCs, corporations … organizations etc… from public, private politics, religions, sciences, governments, etc… This is ancestors facts Black people are the organic humans with blood 🩸 in their bodies no mattter the skin shades… black comes from the word blood 🩸 Blood = Black We are the people of color because we have so much different colors of skin shades What this means is that people with pale skin are black and people of color (they are not white) universally coming from Africa or not PERIOD! Education is the key 🔐 that open doors 🚪 In this photo you can see a white and black human in one photo ❤️

Lol. This is the best reply from a person who actually knows what AI is: a scientist.

@ManicMindTrick

16 күн бұрын

1. You can find a large number of people working on the cutting edge of AI with the direct goal of making an AGI or have contributed greatly to the field that are very cognisant and verbal about AI catastrophic risks. Whether those noises are strategic and result in actual AI safety is another thing but a lot of leading AI people are deeply concerned. 2. Don't trust a man whose salary depends on him not understanding. 3. The arguments he proposes here is either just vague or insanely naive, like "Just shutting it off". When I hear that response I know they haven't thought about this issue at all or heard any counterargument ever. Nice one Einstein. At least someone like Neil Degrasse Tyson understood the stupidity of that argument in time and changed his mind.

Yeah OK, Extinction Level Risk might be pushing it a bit, however; 5:12 With machine learning, once the program is given the ability to rewrite itself, even the people who wrote it are sometimes unable to determine exactly how the program reaches a decision. His argument of 'Oh, we can just rewrite the code' is a gross oversimplification at best and would not necessarily prevent a repeat of the prior fault in any case. 5:50 Actually I do think we need to consider AI as all these things because it's with the integration of different systems that the greatest risk arises. We've seen already how easy it is to get software such as Chat GPT to break it's own rules, so what happens when one branch of AI is developed and used to interface with another? It is ridiculously naive to suggest that we can come up with some foolproof unbreakable algorithm to prevent errors in just one of those systems never mind one AI interacting with another, particularly in view of my comment above regarding machine learning. In fact his attitude here is one common way that bugs slip through QA, i.e. thorough testing of individual functions in isolation but not enough time testing the whole thing. 7:10 I'm sorry, did this guy really just say that a computer program needs intent to function differently than intended? Yes. Yes he does go on to say that AI would need to have programmed harmful intent to pose a risk. OK, fine let's just blithely press on then and find out what the problems are when we get there. Isn't that what mankind does best?

@charlesmanning3454

4 ай бұрын

1. "Once the program is given ability to rewrite itself" ... Rewrite itself how? I don't mean the mechanism by which it will generate new code but how will it evaluate the new code? A.I. can write code but it doesn't invent the requirements for the code. The aim or purpose or requirements always come from outside the A.i. 2. The is always a danger that programs, machines or systems will not function how we expect them to. Create an A.I. to control some system and maybe the A.I. will malfunction and the system will fail. Create a procedural program to control some system and maybe the program will malfunction and the system will fail. Complex systems have risks due to complexity. There no reason to believe a malfunction A.I. will learn to become better and better at malfunctioning. 3. No, he said computers don't have intent. We shouldn't think they or talk as if they do. A machine created to make paper clips may not make paper clips as intended but it's not going to take that indent and run with it.

@ManicMindTrick

16 күн бұрын

@@charlesmanning3454 1. It will create a virtual machine and run and evaluate code running simulation. 2. Systems like ChaptGPT are black boxes and we don't know how they operate. This will become more and more apparent and potentially dangerous over time. 3. Computers have goals and sub-goals not intent. It's easy to anthropomorphize AI, especially as it is trained on human data, uses human language and even speaks to us using a human like voice but it is a completely alien type of intelligence.

@ManicMindTrick

16 күн бұрын

Why is extinction risk "pushing it"? I never heard a decent argument so far and I have heard a lot of people trying their best to push that narrative. My bad if I sound condescending but hearing people being so cavalier about potential existential risk is my definition of really pushing it.

@bassmaiasa1312

Күн бұрын

​@@charlesmanning3454 When he says a computer would need to form an intent to wipe out humanity, isn't he anthromorphizing the question? Does the COVID virus have an intent to kill its host? A comet would not need an intent to hit the Earth. The fact that computers don't form human motives is not reassuring in the slightest. On the contrary, the reason the same species that killed 70 million people in WWII did not take the next step to thermonuclear extinction (so far) is we never formed the intention to kill ourselves off completely.

This gentleman is QUESTIONING "SCIENCE"??? 👍 I pray for his safety and that of his family. "They" won't take kindly to this. I APPLAUD 👏 HIM.

@ManicMindTrick

16 күн бұрын

I have no idea what you just wrote.

It seems like a lot of the comments on this debate are AI-generated. I know comments have trended in the direction of becoming shorter and more stupid over time so it's hard to tell sometimes.

@DanielMoreno-ih2cy

10 күн бұрын

One of the many future threats of AI

@DanielMoreno-ih2cy

10 күн бұрын

Also KZread is full of ininteligible morons

What he know when he is dead and gone?. Shortly

From 10 minutes around i started seeing the greatness of AI and the world that is incombent upon us.

@rajeevdsamuel

3 ай бұрын

You are just a chatbot virgin.

The most well educated person on the planet might sound crazy on hallucinogenic drugs, unlike AI, at least the most educated person can "feel" others pain via internalized conscious deliberation...AI can't and probably never will.