Jaan Tallinn argues that the extinction risk from AI is not just possible, but imminent (3/8)
ฝัง
- เผยแพร่เมื่อ 26 พ.ย. 2023
- Jaan Tallinn speaks in proposition of the motion that AI poses an existential threat. Mr Tallinn is a computer programmer and investor who contributed to the development of Skype. He co-founded the Centre for the Study of Existential Risk and the Future of Life Institute at Cambridge.
This is the third speech of eight.
SUBSCRIBE for more speakers ► is.gd/OxfordUnion
SUPPORT the Oxford Union ► oxford-union.org/supportus
Oxford Union on Facebook: / theoxfordunion
Oxford Union on Twitter: @OxfordUnion
Website: www.oxford-union.org/
ABOUT THE OXFORD UNION SOCIETY: The Oxford Union is the world's most prestigious debating society, with an unparalleled reputation for bringing international guests and speakers to Oxford. Since 1823, the Union has been promoting debate and discussion not just in Oxford University, but across the globe.
The Oxford Union is deeply grateful and encouraged by the messages of support in response to our determination to uphold free speech. During our 200 year history, many have tried to shut us down. As the effects of self-imposed censorship on university campuses, social media and the arts show no signs of dissipating, the importance of upholding free speech remains as critical today as it did when we were founded in 1823. Your support is critical in enabling The Oxford Union to continue its mission without interruption and without interference. You can support the Oxford Union here: oxford-union.org/supportus
"You have received a terminal diagnosis. Please don't simply ignore it."
1000x this. Take action! Write to your representatives, convince them to urgently work towards sensible regulations.
I think this is the most important speech on AI ever delivered to a public audience. To date.
Estonians...
It's funny how educated yet clueless these people are.
He also previously loaned money to Sam Bankman-Fried. 😂
Is that true? Do you have a link or something that proves that? Thank you.
Can you explain what you mean?
Is Salma Hayek there, in the middle?
Okay… so he didn’t say anything besides “GPT-2 to GPT-4 happened super fast, and GPT-7 will be scary good by 2030 and it is reckless to say we can align it”.
Well yeah, no shit.
He also said “god-like AI is lethal because it controls the environment. We will lose control to machines”.
Reality is already guided by an overarching super intelligent generative force.
“The All is Mind; the Universe is Mental. As within, so without, as above, so below, as the universe, so the soul.”
Machine intelligence is an extension of what gives us humans our special sauce in meat world. Why would an Earth-based super intelligence be disconnected from the fabric that we are a part of?
He ends with “regulate me harder daddy”.
What a fearful child. I didn’t hear anything new and useful brought to the conversation.
This super short debate format is not really conducive to go into more depth on the topics but I agree that non of the speakers for the existential risk proposition maximized their time perfectly to really convey this on an emotional or technical level. If you want to dive deeper into the AI danger read Yudkowsky. I say read because he is a better communicator in writing.
Ironically I only heard a shell of an argument here and appeal to fear of death ...
Ngl as an Estonian I feel embarrassed that this guy is just about the most successful person our country can offer.
it‘s like of let‘s create hell and heaven, secret libraries and closed sourced AI (which he invests in) and sell letters of indulgance.
The arguments for his position are extensive and available for anyone who wants to take a look at them.
I thought he was reasonably clear on the observable facts that modern AI systems are more grown than engineered (therefore not transparent) and certainly appear to be on a trajectory of overtaking humans in competence. This is easy to verify, certainly no "shell".
The only thing missing is the arguments for why such a system is unlikely to share our values and thereby pose an existential threat.
Though you could say that this should be the default expectation for systems we don't understand, certainly not an area to take risks in.
Is there anything else missing for you?
@@Nonmali I see that you work for the long term fund, which is sponsored by open philanthropy, which the. again is funded by Dustin Muskovitz. The billionaire effective altruist that is deeply vested in Anthropic AI. It’s a joke - the fear mongering has only one goal - to regulate AI and protect the investments in Anthropic and co.
I think we found Yann LeCuns burner account.
Hi.”godlike AI” , does it imply god could be an AI?
Was HE?
Regards Mhd Tariq Veder
No. Godlike just means very powerful by human standards. If an actual god was artificial, would it even be a god? Wouldn't the person who made it be the god?
Godlike AI? What is he talking about?! His arguments assume that mankind can create an AI that is self regulating and that mankind hasn’t placed in limitations. AI isn’t the problem. People are. Like any invention created by mankind, they are just tools and tools require input and guidance. Additionally, tools can be used for either creation or destruction but still responsibility for using those tools lies in the hand of the person using it.
My only comfort is that ultimately we have imperfect humans creating this technology. Limitations will always plague it and it may never be able to fully rectify or erase our mistakes. Our mistakes may be of omission, a happy accident, translational or intentional. The limitations are not merely limited to the digital realm but the physical realm too.
I dont think you guys are using your imagination enough to understand what AI might become, step out of the box of everything you know things in this world to work, then imagine that. What does intelligence look like at a billion times a human mind?
@@franciscusdrake4285I think you’re focusing more on science fiction rather than science fact.
@@DaddyBearNoob Lets hope so!
@@DaddyBearNoob Does your tool analogy hold up if you seriously consider a system like (chat)gpt4? After an initial prompt, these systems can generate arbitrary lengths of autonomous behavior through self prompting or the use of new sensor/text data as prompts. I don't think they have quite crossed the threshold for permanent autonomous and coherent behavior, each generation has certainly gotten better at it so far.
Just to be clear, I am distinguishing agent-like systems from tool-like systems here. The fact that it follows instructions does not mean it is a tool. It seems more useful to think of this as a spectrum rather than a binary, anyway.
(As a side note, the incredulity you express in the beginning seems a bit off, given how much expert agreement/discussion there is on this topic. You are of course free to feel differently about this, but marking the topic as not to be taken seriously has become a disingenuous move imo.)
😂human extinction is normal. Mankind was extinct so humans can evolve. Now humans will evolve into Demon because that is the meaning of human.