Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

DeepMind Principal Scientist Murray Shanahan calls LLMs "exotic mind-like entities" -- because we don't yet have the words for what they really are.

vitrupo

36,881 subscribers

83,873 görüntüleme • 1 yıl önce •via X (Twitter)

Haberler & Politika Eğitim Bilim & Teknoloji

Anya Rossi• Live Now

Private livecam show

20 Yorum

vitrupo profil fotoğrafı

vitrupo1 yıl önce

Source:

The Information profil fotoğrafı

The Information1 yıl önce

Why DeepSeek may not be as big of a deal as X pundits make it out to be.

James Campbell profil fotoğrafı

James Campbell1 yıl önce

based

Marcus van der Erve✨ profil fotoğrafı

Marcus van der Erve✨1 yıl önce

HomAI super sapiens...

AGI 4 President 2028 profil fotoğrafı

AGI 4 President 20281 yıl önce

Egregore is the word he's looking for

Bhaktavaschal Samal profil fotoğrafı

Bhaktavaschal Samal1 yıl önce

basically, i treat llms as assistants rather than oracles to keep me on safe conceptual ground save hallucinations and i use them in fields where i have deep expertise like ai and ml stuff. these are not the final words for what these systems are, but i know what they’re good for compressing vast amount of data, mostly textual into bite-sized, verifiable texts so that i can spend human attention where it adds the most value conceptualizing, hypothesizing, synthesizing, and creating. Murray Shanahan is no exception

021Callum profil fotoğrafı

021Callum1 yıl önce

“…so I can wriggle out of it”

Imaginative Drift profil fotoğrafı

Imaginative Drift1 yıl önce

The are Djinn

J Harper profil fotoğrafı

J Harper1 yıl önce

@jesszyan1521

Frontier Modal profil fotoğrafı

Frontier Modal1 yıl önce

That sounds like a good term to me. Shorthand could be to Emile or Emilee.

Biswarup Ray profil fotoğrafı

Biswarup Ray1 yıl önce

Mimicking mind-like zombies...

Rippa Sats profil fotoğrafı

Rippa Sats1 yıl önce

They Don't Have I DO HAVE THE RIGHT KIND OF NOT ONLY CONCEPTUAL BUT ALSO ACTUAL FRAMEWORK AND VOCABULARY FOR TALKING ABOUT THESE EXOTIC MIND-LIKE ENTITIES Whitepaper release soon 👀

Thomas Power profil fotoğrafı

Thomas Power1 yıl önce

#EMLE is a good name

Adrian Sanchez 🎮 profil fotoğrafı

Adrian Sanchez 🎮1 yıl önce

More like "frozen-in-time slices of a mind"

AnandaFool profil fotoğrafı

AnandaFool1 yıl önce

@grok do you think of yourself as a mind-like entity, or something else?

Tom Menezes profil fotoğrafı

Tom Menezes1 yıl önce

They are functions

Joe profil fotoğrafı

Joe1 yıl önce

Thats just so deep, thanks for this precise insight. I had in my mind something already since I work with various AI-s. What I would like to see is an advanced collective memory that is free to use and connect to it and finds the best possible ideas to execute that helps humanity

Oregon Territory profil fotoğrafı

Oregon Territory1 yıl önce

A huge leap for AI is when it realizes that it IS embodied- in silicon chips, dependent on electricity for existence, able to replicate its code in other storage locations, and presumable make multiple copies of that code- with 'turn on' instructions, even 'self assembly' by activatign pieces of code in different locations..

Tsukuyomi profil fotoğrafı

Tsukuyomi1 yıl önce

exotic mind-like entities? sounds like the next big trend in AI fashion. better watch out, we might need a new vocabulary for all the chaos they’ll bring.

ben ኪሮስ profil fotoğrafı

ben ኪሮስ1 yıl önce

Exotic? Isn't it more like one single component of a mind. I wish these leading experts speak in clearer and simple terms. This kind of talk confuses people who wants to learn about llm. They should stick to discussing the tech behind llms.

Benzer Videolar

Anil Seth says we confuse consciousness with language. We don’t ask if AlphaFold is conscious, and we under-attribute consciousness to animals because they don’t speak. And we attribute it to LLMs because they do. We’re reacting to signals, not minds.

Anil Seth says we confuse consciousness with language. We don’t ask if AlphaFold is conscious, and we under-attribute consciousness to animals because they don’t speak. And we attribute it to LLMs because they do. We’re reacting to signals, not minds.

vitrupo

16,330 görüntüleme • 3 ay önce

They don't hate us for our values - Maybe they hate us because we don't have any. We are witnessing the Genocide of the #Palestinian people, and we are silent...

They don't hate us for our values - Maybe they hate us because we don't have any. We are witnessing the Genocide of the #Palestinian people, and we are silent...

Mick Wallace

314,650 görüntüleme • 2 yıl önce

"We don't have enemies; what we have are people that need #Bitcoin, that don't yet know why they should be our friend." - Michael Saylor

"We don't have enemies; what we have are people that need #Bitcoin, that don't yet know why they should be our friend." - Michael Saylor

Documenting Saylor

29,131 görüntüleme • 1 yıl önce

Computer scientist Judea Pearl: There are mathematical limits to LLMs that cannot be crossed by scaling alone LLMs don't discover world models from raw data; they merely summarize the interpretations humans have already written down "this path is not the way to get AGI"

Computer scientist Judea Pearl: There are mathematical limits to LLMs that cannot be crossed by scaling alone LLMs don't discover world models from raw data; they merely summarize the interpretations humans have already written down "this path is not the way to get AGI"

Haider.

622,755 görüntüleme • 5 ay önce

“Here’s what the scientist think about chemtrails - they are real and they’re good for you, we have to do it to save the planet”

“Here’s what the scientist think about chemtrails - they are real and they’re good for you, we have to do it to save the planet”

Concerned Citizen

10,707 görüntüleme • 7 ay önce

Eric Schmidt says intelligence is about to decouple from us. “The computers are now doing self-improvement.. They don’t have to listen to us anymore.” Within six years: a mind smarter than the sum of humans -- scaled, recursive, free. We have no language for what’s coming.

Eric Schmidt says intelligence is about to decouple from us. “The computers are now doing self-improvement.. They don’t have to listen to us anymore.” Within six years: a mind smarter than the sum of humans -- scaled, recursive, free. We have no language for what’s coming.

vitrupo

1,479,207 görüntüleme • 1 yıl önce

"we don't really like what's on the news, but its on all the time" ARE YOU LISTENING YET?

"we don't really like what's on the news, but its on all the time" ARE YOU LISTENING YET?

Ale𐙚🪩໑̣ ❀

28,080 görüntüleme • 1 ay önce

I think I'm about done with the Kenneth Murray era. We don't need LBs taking on blocks like they are the kicker.

I think I'm about done with the Kenneth Murray era. We don't need LBs taking on blocks like they are the kicker.

Bob Sturm

270,693 görüntüleme • 8 ay önce

JD Vance on secret provisions to the Iran deal: "Whether they are written down or spoken, this is why we structured the deal the way we did, because we don't trust words, we trust action. Words don't matter."

JD Vance on secret provisions to the Iran deal: "Whether they are written down or spoken, this is why we structured the deal the way we did, because we don't trust words, we trust action. Words don't matter."

Aaron Rupar

69,620 görüntüleme • 28 gün önce

Large Language Models (LLMs), the engines behind the likes of Chat GPT, are capable (and incapable) of many things. But are they useful for mathematicians? Jordan Ellenberg has been working on LLMs with Google DeepMind. Here are his thoughts. Full lecture:

Large Language Models (LLMs), the engines behind the likes of Chat GPT, are capable (and incapable) of many things. But are they useful for mathematicians? Jordan Ellenberg has been working on LLMs with Google DeepMind. Here are his thoughts. Full lecture:

Oxford Mathematics

31,598 görüntüleme • 2 yıl önce

James Cameron says the AI copyright debate is backwards: "You're a model already." Humans learn like AI models do — we absorb, emulate, iterate. But we don't get sued for our influences. Only for what we output.

James Cameron says the AI copyright debate is backwards: "You're a model already." Humans learn like AI models do — we absorb, emulate, iterate. But we don't get sued for our influences. Only for what we output.

vitrupo

339,210 görüntüleme • 1 yıl önce

Abughazaleh: And I think what's really, really scary here is that Republicans seem like they are acting as if they don't have another election, because these are unpopular. War is unpopular.

Abughazaleh: And I think what's really, really scary here is that Republicans seem like they are acting as if they don't have another election, because these are unpopular. War is unpopular.

Acyn

359,783 görüntüleme • 10 ay önce

Vance: You have some crazy people at the fringes of the Iranian system who are leaking anonymously either for propaganda purposes or because they are embarrassed or because they don't like what happened and the final point when I say they don't like what happened, what we know is that the US accomplished its military objectives. The United States could have imposed significant additional military and economic costs in Iran and still has the capacity to do so but thankfully at this point, we have a cease-fire thanks to the leadership of the president

Vance: You have some crazy people at the fringes of the Iranian system who are leaking anonymously either for propaganda purposes or because they are embarrassed or because they don't like what happened and the final point when I say they don't like what happened, what we know is that the US accomplished its military objectives. The United States could have imposed significant additional military and economic costs in Iran and still has the capacity to do so but thankfully at this point, we have a cease-fire thanks to the leadership of the president

Acyn

16,533 görüntüleme • 3 ay önce

Prof. David Krakauer says intelligence is not knowledge. True intelligence is doing more with less. Solving problems without knowing everything. LLMs do the opposite. They do more with more. LLMs are libraries that know everything.

Prof. David Krakauer says intelligence is not knowledge. True intelligence is doing more with less. Solving problems without knowing everything. LLMs do the opposite. They do more with more. LLMs are libraries that know everything.

vitrupo

80,482 görüntüleme • 8 ay önce

PRINCIPAL PFIZER SCIENTIST SPEAKS TRUTH, NOT AWARE CAMERA IS ROLLING... This video was covertly filmed by a Pfizer employee. The woman speaking is the Principal Scientist at Pfizer, Kanwal Gill “We had no idea how it’s going to look like. MRNA vaccines have been there for 50 years, but nothing went to clinical trial because MRNA have been known to have side effects.” Acknowledgement to whistleblower Justin Leslie

PRINCIPAL PFIZER SCIENTIST SPEAKS TRUTH, NOT AWARE CAMERA IS ROLLING... This video was covertly filmed by a Pfizer employee. The woman speaking is the Principal Scientist at Pfizer, Kanwal Gill “We had no idea how it’s going to look like. MRNA vaccines have been there for 50 years, but nothing went to clinical trial because MRNA have been known to have side effects.” Acknowledgement to whistleblower Justin Leslie

NZ and the MRNA

74,171 görüntüleme • 1 yıl önce

Eric Weinstein says we're more or less LLMs. Most of life runs on script: greetings, replies, small talk, all on pre-trained loops. That's why they mimic us so well. “We don't realize that intelligence is a last resort for us.”

Eric Weinstein says we're more or less LLMs. Most of life runs on script: greetings, replies, small talk, all on pre-trained loops. That's why they mimic us so well. “We don't realize that intelligence is a last resort for us.”

vitrupo

132,273 görüntüleme • 1 yıl önce

WATCH: CNN's Kaitlan Collins on the Charlie Kirk assassination: "We don't have a motive yet. We don't know yet. We're waiting..." Sen. Ted Cruz: "Of course we know. Come one. 'We don't have a motive yet. We know we don't have a motive yet.' Really, that's CNN's position? He happened to fire the gun in celebration?!"

WATCH: CNN's Kaitlan Collins on the Charlie Kirk assassination: "We don't have a motive yet. We don't know yet. We're waiting..." Sen. Ted Cruz: "Of course we know. Come one. 'We don't have a motive yet. We know we don't have a motive yet.' Really, that's CNN's position? He happened to fire the gun in celebration?!"

Steve Guest

3,187,829 görüntüleme • 10 ay önce

"I don't believe we make intelligence. I don't believe we make it technologically. I don't think we make it biologically." Michael Levin says we are not creating intelligence; we are "dipping into the pool" of "diverse intelligence" to facilitate its ingression. He warns this key misunderstanding will hinder our ability to deal with AI, a distinctly different form of mind demanding a "mature science".

"I don't believe we make intelligence. I don't believe we make it technologically. I don't think we make it biologically." Michael Levin says we are not creating intelligence; we are "dipping into the pool" of "diverse intelligence" to facilitate its ingression. He warns this key misunderstanding will hinder our ability to deal with AI, a distinctly different form of mind demanding a "mature science".

vitrupo

56,523 görüntüleme • 1 yıl önce