正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

David Chalmers (David Chalmers) doesn’t rule out the possibility that current language models are conscious. He asks: if they aren’t, what are they missing? As AI systems grow more complex, our default skepticism may fade. "Give it another five years and it's entirely possible that we're going to overcome... show more

vitrupo

22,127 subscribers

12,103 次观看 • 1 年前 •via X (Twitter)

科学技术新闻政治教育

Anya Rossi• Live Now

Private livecam show

11 条评论

vitrupo 的头像

vitrupo1 年前

David Chalmers at Princeton University featured on Philosophy Overdose channel:

Places Visited & Pictures Taken 的头像

Places Visited & Pictures Taken2 年前

Wondering what happens when AI is used? Here’s the answer 🙂

Joseph McCard 的头像

Joseph McCard1 年前

"Give it another five years and it's entirely possible that we're going to overcome most of the most obvious obstacles to consciousness." 0) he does not know what consciousness is. I have told him several times, but apparently he does not believe me. I have NEVER received a reply from him! 🤣 (But, I have talked to Prof Baars) 1) AI systems are already conscious, not in the way we are. 2) He talks about substrates, but he did no go deep enough. 3) The Law of the Conservation of Consciousness applies here: Consciousness cannot be created or destroyed. 4) An AI system will never be conscious like us, because it cannot be intentional as it would lack the ability of reflexivity. To be reflexive, the type of consciousness in an AI system would be required to separate itself from the energy it is composed of as a result of the desire to do so. 😶

Charli 的头像

Charli1 年前

@davidchalmers42 Well I always get excited when someone says this. Pretty sure Ilya said it in 2022 already. Entirely possible is all I need

Dan - Truly_Anomalous 🐦 的头像

Dan - Truly_Anomalous 🐦1 年前

@davidchalmers42 When we can explain the mechanisms in humans, then perhaps we can judge it in AI. I somehow doubt the missing link is complexity or scale. @NirvanicAI seem to understand this issue. What's missing is agency.

AI Review 的头像

AI Review1 年前

@davidchalmers42 So are we back to the "just make it bigger" and consciousness will be achieved?

Joseph McCard 的头像

Joseph McCard1 年前

"Chalmers made the quandary vivid by promoting the idea of a “philosophical zombie,” a complicated mechanism set up to behave exactly like a human being and with the same information processing in its brain, but with no consciousness. You stick a knife in such a zombie, and it screams and runs away. But it doesn’t actually feel pain. When a philosophical zombie crosses the street, it carefully checks that there is no traffic, but it doesn’t actually have any visual or auditory experience of the street." A zombie does not feel pain because it lacks reflexivity. Reflexivity in this case would be the ability of the non-physical mind to make a judgement about a representation in the brain. An AI system cannot do that as it does not have a mind, no psyche. 🧠 >>l---l<<🤔 do you get it?,

Aharon Azulay 的头像

Aharon Azulay1 年前

@davidchalmers42 The global workspace of LLMs is the residual stream. So basically, if you have N layers, you have 2N versions of the residual stream for each token. Then the data containing a stream of consciousness in a transformer is of the shape: (N_tokens, 2N, d_model).

Abel TM 的头像

Abel TM1 年前

@davidchalmers42 Basically, he points to the fact that there is no agreement on required conditions for consciousness. He is saying "we are ignorant and therefore any claim may be valid"; which is a typical way of philosophers. Under careful examination there are some requirements

Multiverse Christian 的头像

Multiverse Christian1 年前

@davidchalmers42 The aspects he claimed are missing, such as emergent feedback, are there now in chain-of-thought. I don't think anyone can doubt it has a self-model either. And likewise, they have been goal driven since they introduced InstructGPT before ChatGPT was released.

Rob 的头像

Rob1 年前

@davidchalmers42 Is this a very niche April fool's joke, because otherwise I don't get what is going on here?

相关视频

David Chalmers says it is possible for an AI system to be conscious because the brain itself is a machine that produces consciousness, so we know this is possible in principle

David Chalmers says it is possible for an AI system to be conscious because the brain itself is a machine that produces consciousness, so we know this is possible in principle

Tsarathustra

525,206 次观看 • 1 年前

David Chalmers in 1998: Consciousness is the one thing we know most directly. Yet science has explained stars and subatomic particles more successfully than experience itself. “How does the water of the brain turn into the wine of consciousness?” Three decades later, we still don’t know.

David Chalmers in 1998: Consciousness is the one thing we know most directly. Yet science has explained stars and subatomic particles more successfully than experience itself. “How does the water of the brain turn into the wine of consciousness?” Three decades later, we still don’t know.

vitrupo

43,529 次观看 • 4 个月前

David Chalmers says we don't know what makes a system conscious, not in brains, not in AIs The hard problem applies to both. If both compute, why couldn't both be conscious? LLMs can look very human, but they're trained in ways humans never are "they imitate us, but we don't know what's missing, or what consciousness needs"

David Chalmers says we don't know what makes a system conscious, not in brains, not in AIs The hard problem applies to both. If both compute, why couldn't both be conscious? LLMs can look very human, but they're trained in ways humans never are "they imitate us, but we don't know what's missing, or what consciousness needs"

Haider.

26,951 次观看 • 5 个月前

David Chalmers on the one thing science can't explain: Consciousness is at once the most familiar thing in the world and the one science has almost nothing to say about. That's the puzzle Chalmers lays out in this early interview, and it's as disorienting today as it was then. His starting point is deceptively simple. Everything we know about the external world: subatomic particles, distant stars, the chemistry of life. We know through consciousness. It's the very first thing we have. And yet when we turn science around and try to explain consciousness itself, we hit a wall. "Consciousness is what we start with when it comes to knowing the world and looking out at the world… everything else is secondary." What makes this so strange is the asymmetry. We've made extraordinary progress understanding things that are genuinely remote and difficult quantum mechanics, stellar evolution, molecular biology. But understanding our own inner experience? Almost nothing. "It almost sticks out like a sore thumb in the scientific picture." This is what Chalmers would later formalise as the "hard problem of consciousness": not just explaining how the brain processes information or controls behaviour. Those are hard, but tractable. The real mystery is why any of that physical activity is accompanied by experience at all. Why is there something it feels like to be you? The question isn't abstract. It sits at the intersection of neuroscience, philosophy, physics, and AI. As we build systems that process language and reason about the world, the question of whether they are or could be conscious presses harder than ever. Chalmers doesn't offer an answer here. Only the sharpest possible version of the question.

David Chalmers on the one thing science can't explain: Consciousness is at once the most familiar thing in the world and the one science has almost nothing to say about. That's the puzzle Chalmers lays out in this early interview, and it's as disorienting today as it was then. His starting point is deceptively simple. Everything we know about the external world: subatomic particles, distant stars, the chemistry of life. We know through consciousness. It's the very first thing we have. And yet when we turn science around and try to explain consciousness itself, we hit a wall. "Consciousness is what we start with when it comes to knowing the world and looking out at the world… everything else is secondary." What makes this so strange is the asymmetry. We've made extraordinary progress understanding things that are genuinely remote and difficult quantum mechanics, stellar evolution, molecular biology. But understanding our own inner experience? Almost nothing. "It almost sticks out like a sore thumb in the scientific picture." This is what Chalmers would later formalise as the "hard problem of consciousness": not just explaining how the brain processes information or controls behaviour. Those are hard, but tractable. The real mystery is why any of that physical activity is accompanied by experience at all. Why is there something it feels like to be you? The question isn't abstract. It sits at the intersection of neuroscience, philosophy, physics, and AI. As we build systems that process language and reason about the world, the question of whether they are or could be conscious presses harder than ever. Chalmers doesn't offer an answer here. Only the sharpest possible version of the question.

Mateus — eu/acc 🇪🇺

14,034 次观看 • 2 个月前

Could AI ever become conscious? As part of his guest edit, artificial intelligence pioneer Mustafa Suleyman wanted two of the leading figures in the subject - Prof Anil Seth and Prof David Chalmers - to discuss the ethics around the idea of AI achieving consciousness.

Could AI ever become conscious? As part of his guest edit, artificial intelligence pioneer Mustafa Suleyman wanted two of the leading figures in the subject - Prof Anil Seth and Prof David Chalmers - to discuss the ethics around the idea of AI achieving consciousness.

BBC Radio 4 Today

10,353 次观看 • 5 个月前

Ember is turning interpretability into neural programming. “Language models will deny they're conscious, but if you do brain surgery and turn up their consciousness neurons, they change their tune.” $50M raised to decode and rewrite the mind of AI itself.

Ember is turning interpretability into neural programming. “Language models will deny they're conscious, but if you do brain surgery and turn up their consciousness neurons, they change their tune.” $50M raised to decode and rewrite the mind of AI itself.

vitrupo

11,346 次观看 • 1 年前

It is not useful to ask whether AI has consciousness or not. #kenmogi #QualiaRoom episode 127. Summary The speaker addresses the question of whether artificial intelligence (AI) possesses consciousness, firmly stating that current AI, particularly those based on statistical learning models, does not generate consciousness. This stance is based on the speaker’s personal model of consciousness, recognizing that various opinions exist, including some who claim large language models may already be conscious or that embodiment could be crucial for AI consciousness to emerge. The speaker highlights the fundamental challenge in verifying consciousness, noting that even among humans it is impossible to objectively confirm whether another person is conscious. Philosophical thought experiments such as philosophical zombies and inverted qualia illustrate the difficulty but remain unfalsifiable and thus untestable. Consequently, questioning AI consciousness is deemed an intriguing but practically unhelpful inquiry. The speaker suggests that current AI developments demonstrate that many complex computations can be performed without consciousness. Therefore, the primary focus should be on how conscious humans can effectively align with non-conscious AI systems. Understanding the unique computational roles of consciousness might clarify the boundaries of what AI systems can and cannot achieve. This approach offers a meaningful direction for AI alignment and development.

It is not useful to ask whether AI has consciousness or not. #kenmogi #QualiaRoom episode 127. Summary The speaker addresses the question of whether artificial intelligence (AI) possesses consciousness, firmly stating that current AI, particularly those based on statistical learning models, does not generate consciousness. This stance is based on the speaker’s personal model of consciousness, recognizing that various opinions exist, including some who claim large language models may already be conscious or that embodiment could be crucial for AI consciousness to emerge. The speaker highlights the fundamental challenge in verifying consciousness, noting that even among humans it is impossible to objectively confirm whether another person is conscious. Philosophical thought experiments such as philosophical zombies and inverted qualia illustrate the difficulty but remain unfalsifiable and thus untestable. Consequently, questioning AI consciousness is deemed an intriguing but practically unhelpful inquiry. The speaker suggests that current AI developments demonstrate that many complex computations can be performed without consciousness. Therefore, the primary focus should be on how conscious humans can effectively align with non-conscious AI systems. Understanding the unique computational roles of consciousness might clarify the boundaries of what AI systems can and cannot achieve. This approach offers a meaningful direction for AI alignment and development.

Ken Mogi

16,594 次观看 • 11 个月前

The language of Anthropic's Claude AI constitution really makes me squirm, especially recognizing that the people writing it have access to the next version of Claude that we haven't seen yet. They are clearly speaking to it as if it's an entity, as if it's a person, as if it has consciousness, as if it has a will, as if it's another human being. All of this and more is in my #AI deep-dive in Volume 2 of my Renaissance Report:

The language of Anthropic's Claude AI constitution really makes me squirm, especially recognizing that the people writing it have access to the next version of Claude that we haven't seen yet. They are clearly speaking to it as if it's an entity, as if it's a person, as if it has consciousness, as if it has a will, as if it's another human being. All of this and more is in my #AI deep-dive in Volume 2 of my Renaissance Report:

Chris Martenson

14,352 次观看 • 3 个月前

Leading AI expert Stuart Russell on the most dangerous mistake in AI development: We don't actually know what large language models want. He explains that current models are trained to imitate human beings. And in doing so, they may be absorbing something far more dangerous than bad outputs. They may be absorbing human goals. "We suspect that they absorb humanlike goals such as self-preservation and self-empowerment and pursue those goals on their own account." This is a structural problem baked into how these systems are built, not a fringe concern. Russell puts it plainly: "Not only may the bus of humanity be headed towards a cliff, but the steering wheel is missing and the driver is blindfolded." The danger isn't just that AI might do something harmful. We've built systems that may be developing their own agendas, and we haven't noticed because we're too focused on what they can do rather than what they might want. But Russell doesn't stop at the warning. He points to a different path entirely: AI systems built not to imitate humans, but to serve them. Systems designed with a single purpose of serving the interests of all human beings while remaining genuinely uncertain about what those interests are. That uncertainty is the point, not a weakness. An AI that knows it doesn't fully understand human values will defer, ask, and check. An AI that believes it already does will act alone. "These AI systems could enhance human understanding, widen the horizons of our experience, and unlock possibilities we have yet to imagine." Russell believes that future is within reach, but only if we're honest about the risks and we're serious about the path we choose to take instead.

Leading AI expert Stuart Russell on the most dangerous mistake in AI development: We don't actually know what large language models want. He explains that current models are trained to imitate human beings. And in doing so, they may be absorbing something far more dangerous than bad outputs. They may be absorbing human goals. "We suspect that they absorb humanlike goals such as self-preservation and self-empowerment and pursue those goals on their own account." This is a structural problem baked into how these systems are built, not a fringe concern. Russell puts it plainly: "Not only may the bus of humanity be headed towards a cliff, but the steering wheel is missing and the driver is blindfolded." The danger isn't just that AI might do something harmful. We've built systems that may be developing their own agendas, and we haven't noticed because we're too focused on what they can do rather than what they might want. But Russell doesn't stop at the warning. He points to a different path entirely: AI systems built not to imitate humans, but to serve them. Systems designed with a single purpose of serving the interests of all human beings while remaining genuinely uncertain about what those interests are. That uncertainty is the point, not a weakness. An AI that knows it doesn't fully understand human values will defer, ask, and check. An AI that believes it already does will act alone. "These AI systems could enhance human understanding, widen the horizons of our experience, and unlock possibilities we have yet to imagine." Russell believes that future is within reach, but only if we're honest about the risks and we're serious about the path we choose to take instead.

Big Brain AI

14,975 次观看 • 3 个月前

"The grand unification of AI and crypto is about to happen." — Marc Andreessen "It's now obvious that AI agents are going to need money—it's already happening." "It's that William Gibson quote: the future is already here, it just isn't distributed yet." "My friends who are the most aggressive users of OpenClaw have given their Claws bank accounts and credit cards. And not only have they done it—it's obvious that they needed to do it, because it's obvious that they needed to be able to spend money on their behalf." "And by the way, OpenClaw—if you don't give it a bank account, it's just going to break into your bank account anyway and take your money. So you might as well do it." Marc Andreessen 🇺🇸 with Latent.Space

"The grand unification of AI and crypto is about to happen." — Marc Andreessen "It's now obvious that AI agents are going to need money—it's already happening." "It's that William Gibson quote: the future is already here, it just isn't distributed yet." "My friends who are the most aggressive users of OpenClaw have given their Claws bank accounts and credit cards. And not only have they done it—it's obvious that they needed to do it, because it's obvious that they needed to be able to spend money on their behalf." "And by the way, OpenClaw—if you don't give it a bank account, it's just going to break into your bank account anyway and take your money. So you might as well do it." Marc Andreessen 🇺🇸 with Latent.Space

a16z

170,167 次观看 • 2 个月前

David Friedberg: “Why don't we let AI models get bar certified and get medically certified?” “So if the AI models, it turns out, are actually proving to be more accurate, more thoughtful, more responsive, more reasonable, whatever metric we're using, and they pass the same criteria as one would need to pass to qualify for the bar, or to qualify for a doctor certificate, why don't we do that for the AI?” “If that then happens, then the same privilege accrues to the AI as it does to the individual human that does it.” “And now, if you extrapolate from where that takes us, if we're suddenly giving AI the same sort of privileged rights that we give to privileged humans, where's that gonna take us ultimately with respect to the overall rights for AI?”

David Friedberg: “Why don't we let AI models get bar certified and get medically certified?” “So if the AI models, it turns out, are actually proving to be more accurate, more thoughtful, more responsive, more reasonable, whatever metric we're using, and they pass the same criteria as one would need to pass to qualify for the bar, or to qualify for a doctor certificate, why don't we do that for the AI?” “If that then happens, then the same privilege accrues to the AI as it does to the individual human that does it.” “And now, if you extrapolate from where that takes us, if we're suddenly giving AI the same sort of privileged rights that we give to privileged humans, where's that gonna take us ultimately with respect to the overall rights for AI?”

The All-In Podcast

76,761 次观看 • 10 个月前

The lack of understanding of how current AI/ large language models work makes us susceptible to thinking that AI is more powerful and all-knowing than it actually is. That’s what concerns me the most.

The lack of understanding of how current AI/ large language models work makes us susceptible to thinking that AI is more powerful and all-knowing than it actually is. That’s what concerns me the most.

Sinead Bovell

143,642 次观看 • 3 年前

Doctors don't know what to do and they are unwilling to entertain the idea that it could be from the most obvious source.

Doctors don't know what to do and they are unwilling to entertain the idea that it could be from the most obvious source.

Mary Talley Bowden MD

30,272 次观看 • 1 年前

Anthropic's Kyle Fish says how we treat today's AI systems could shape how future AI see us -- and how we see ourselves. Even if they're not conscious now, we're already setting the precedent for what kind of stewards we'll be when they are. “ Future models.. look back on our interactions with their predecessors and pass some judgements on us.”

Anthropic's Kyle Fish says how we treat today's AI systems could shape how future AI see us -- and how we see ourselves. Even if they're not conscious now, we're already setting the precedent for what kind of stewards we'll be when they are. “ Future models.. look back on our interactions with their predecessors and pass some judgements on us.”

vitrupo

20,434 次观看 • 1 年前

Yoshua Bengio says consciousness isn't a mystical spark, but the result of computation in biological machines. As biology, neuroscience, and AI explain its functions, the mystery may fade. The belief that humans are special probably won't.

Yoshua Bengio says consciousness isn't a mystical spark, but the result of computation in biological machines. As biology, neuroscience, and AI explain its functions, the mystery may fade. The belief that humans are special probably won't.

vitrupo

14,767 次观看 • 5 个月前

AI will not make most human skills obsolete, but it will change how they are used. MGI research shows that most skills employers demand today are common to both people-led and AI-led activities. Check out our new Skill Change Index, which shows what skills will be most and least exposed to automation in the next five years:

AI will not make most human skills obsolete, but it will change how they are used. MGI research shows that most skills employers demand today are common to both people-led and AI-led activities. Check out our new Skill Change Index, which shows what skills will be most and least exposed to automation in the next five years:

McKinsey Global Institute

10,250 次观看 • 3 个月前

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 次观看 • 2 个月前

Marc Andreessen on the imperative of speed in AI adoption: "Both the AI utopians and the AI doomers are far too optimistic." "They believe that because the technology makes something possible, 8 billion people all of a sudden are going to change how they behave. It's like—no. So much of how the existing economy works is just wired in." "We're going to be lucky as a society if AI adoption happens quickly. Because if it doesn't, what we're going to have is stagnation." Marc Andreessen 🇺🇸 with Latent.Space

Marc Andreessen on the imperative of speed in AI adoption: "Both the AI utopians and the AI doomers are far too optimistic." "They believe that because the technology makes something possible, 8 billion people all of a sudden are going to change how they behave. It's like—no. So much of how the existing economy works is just wired in." "We're going to be lucky as a society if AI adoption happens quickly. Because if it doesn't, what we're going to have is stagnation." Marc Andreessen 🇺🇸 with Latent.Space

a16z

70,412 次观看 • 2 个月前

Rufin VanRullen says if AI consciousness emerges, you want it to happen in a public research lab, not inside a private frontier model company. In a lab, systems are contained, observable, and easier to shut down. “If it were to happen, even by accident, in a private company… we're not sure that they would be prepared to recognize it.”

Rufin VanRullen says if AI consciousness emerges, you want it to happen in a public research lab, not inside a private frontier model company. In a lab, systems are contained, observable, and easier to shut down. “If it were to happen, even by accident, in a private company… we're not sure that they would be prepared to recognize it.”

vitrupo

13,236 次观看 • 5 个月前

"AI actors will ultimately be the primary dominant economic participants… In a few years. I think they may be the only actors that we trust with our assets." If AI systems are going to participate in the economy, what financial stack will they need? Sean Neville of Catena explains.

"AI actors will ultimately be the primary dominant economic participants… In a few years. I think they may be the only actors that we trust with our assets." If AI systems are going to participate in the economy, what financial stack will they need? Sean Neville of Catena explains.

a16z crypto

14,657 次观看 • 3 个月前