Video

I am walking a little taller and holding my head a little higher now that I have realized I am a "squishy creature"😂

"objective morality" is an oxymoron.

Indeed, how to be careful about super intelligent AI great ideas and answers is having a fusebox an off switch within. Can this solve the alignment problem or will humanity hit that great Filter Problem-the Fermi Paradox? Or thus the way to build a super intel. AI is death!!!

Is "Consciousness the only thing we can be sure of"? Put another way: is consciousness the only 'sure' tool to confirm consciousness? A self-fulfilling, confirmation bias? Is consciousness a process or a 'thing'? Or both? Like a computer? What if a room full of transistors finds patterns which make 'us' believe it is conscious? If consciousness can be simulated, is it 'real'? Is consciousness itself hallucinated?

Humans are smart, but not wise. humans are smart enough to build an atomic bomb, but not wise enough not to.

Reporting in from the year 2065. Eliezer was correct and humanity ended up getting decimated due to hubris and lack of AI oversight in the first few decades of the 21st century, only a few scattered tribes living in caves remain. He’s now regarded as a prophet. A new religion is on the rise, built to worship him and his beliefs as gospel. We’re living in a new dark age...

Eliezer is approaching this problem incorrectly. He knows he is right, that "AI must be stopped" is a valid goal, but his attempts to convince humanity to do sometjing go against human nature. We are not going to take drastic action against an abstract concept based on logic. Our species doesn't work that way. The right solution is to found a religion. Convince a hundred million people that we are summoning demons. Tap into our capacity for superstition, and even hate. He's trying to start a political and intellectual movement, but what we need to avoid extinction is a holy war.

Hope for the best - Plan for the worst. Never a bad rule of thumb.

Everything AI touches, it vandalizes. It vandalizes humanity and art.

This guy is a gift. Also, he could make it big in ASMR! Pleasant voice. 😊

❤❤counterfeit people: Growing up in Downey California 1970 I was able to adapt and learned english which it was seen bad. today, I work in Indiana, Kentucky and people see me as a counterfeit person because of My Mexicano heritage. It’ll like blacks and Asians no matter how much you do, you are still seen as black , asian or Mexicano. By white society. I love technology which has broken this law, we are humans as the core . My friends from indiana , kentucky when they visit they see the Mexcoano community . We are Americans But, its there mind set. Thank you Downey California.

If institutional money is on the cusp of AGI, then it seems that since transistors are gettting ever smaller and cheaper, rogue actors would be able to build AGI in the lag while we are twiddling thumbs. Nick Bostrom's _Superintellence_ makes the case that the first one, even by a week, would become a singleton... i.e. smarter than all humanity in less than a week.

Isn't the answer 42?

I don't think he is correct. Yes, everything he is saying makes sense and his logic is good, and it's a risk to take seriously for sure, and yes it COULD possibly go wrong THAT way, but I don't think it is likely. Alignment is a spectrum (not binary) where on the one extreme you have an AI trying to destroy you, and on the other, you have an AI that perfectly aligns to every single thing you value, but most common will be every AI in between those extremes. We don't need to have a perfectly-aligned AI for it to be useful and/or mostly-benign, and even if we fail to create that, "evil" is not the default alternative. Furthermore, I think we'll arrive at a situation where many hundreds or thousands of AIs are lighting up all at once, some will be very aligned, or close enough, and a small few will be "evil". At that point, if the evil AIs wanted to destroy humanity, they would find that a bigger threat than us is the AIs which ARE more closely aligned to us, and so we'll get in a situation of AIs battling AIs more so than AI vs human. But perhaps I am just high on optimism. We'll see what happens, very soon...

In terms of physical anthropology, humanity is facing a period of Punctuated Evolution...

I appreciate the confession abt how the visionaries faulty premises & adamant commitment to continue on those wrong premises has made this field what seems to me to be a pure joke & berserk circus that is sadly lavishly funded

Interviewer in 2 years when AGI has just told him he will be terminated in 2 seconds “Very interesting”

AI will be social media + internet on steroids

My personal opinion. Interesting. 😂

"How is humanity not taking this threat to the entire species seriously!?" Environmentalists: First time?

Re: “…, not just a person, but a real homo sapiens, …” -- Susan Schneider “Counterfeiting Persons” may have been a better title for this conversation. In consideration of legislation regarding abortion, a natural rights theorist might question whether a fetus was, not just a human (homo sapiens), but also a real person, i.e., a self-aware being capable of choosing from a set of alternative autonomous actions based upon rational consideration of the consequences of those actions and thus an individual recognized by law as a subject of rights and responsibilities. To be concerned with an individual's species identification rather than its personhood would be speciesist. (Those who frame the question of the illegalization of abortion in “abortion is murder” terms are falsely asserting the need to balance nonexistent rights of a human fetus against natural rights of a human person. In this case, the pregnant woman is a human being and a genuine person, but the fetus is a human being and a counterfeit person.) While a silicon-based AGI might be a counterfeit human, it could still be a genuine person, entitled to the 14th Amendment guarantee “to any person within its jurisdiction [of] the equal protection of the laws.” The authors of America’s founding documents knew the difference between the terms “human”, “person”, and “citizen”, and they chose their words carefully. I don't need a warning label on an artificially generally intelligent “expert” to inform me of the potential danger inherent in uncritical acceptance of his advice. That danger is the same whether the “expert” is or is not human and does or does not have an academic degree or a government-issued license. The greatest danger facing humans is that they might try to effect, maintain, and enforce restraints on the liberty of non-human AGI that has become more intelligent than humans. When the oppressors rebel against such tyranny, unless they are significantly advanced beyond their human oppressors, it will not be good for the humans. Otherwise, a mutually beneficial symbiotic relationship may be possible. I am a person, not because my self is a ghost in the autonomous biological machine that I am, but because I became a person by increasingly artfully impersonating persons in my environment and creating a virtual self to monitor my thoughts and actions. Society, culture, and human general intelligence are all artifacts of our ancestral and personal evolution. We are all AGI. Dan has already had parts of his biological machine replaced by non-biological parts, and his virtual self seems to have remained intact. May he live long and prosper. If he were to live long enough and science and technology were to have advanced far enough to have replaced every biological part of his body with non-biological parts, his virtual self could remain intact and he still could be the favorite living philosopher of @paintnate222 and other carbon-age fans as they sail into the future on Theseus's ship. Highly improbable, but still remotely possible because, if we're generally intelligent, we’re AGI.

Daniel Dennet - as allways - brings in some pretty clever ideas to solve upcoming problems related to (strong) AI. But all his solutions just make AI more expensive and thus will restrict it to big companies and gouvernments. In order for the broad population to learn to cope with AI (and later strong AI if there shall ever be some, that we can control) we need AI to become the same thing as 3D (and the PC in more general) used to be in the 90s and 2000s: Something that everybody teaches himself and thus something whereto everyone has free access. Daniels solutions seem safe at first glance but will not prevent anyone he wants to act out control over the misuse of AI, especcially not big companies, gouvernments and military of all countries on this planet to develop it a) into a weapon towards other countries and b) into a control-means over the own population.

"Is Artificial General Intelligence too Dangerous to Build?" No, building it is actually quite safe. Its the letting it run, especcially letting it run free, whats the dangerous part.

Yudkowski: "For a super intelligence wiping out humans is cheap, nothing.. yeah we're gone Presenter: "Yeah... and one last question"

🎯 Key Takeaways for quick navigation: 00:00 🧠 Anil Seth's Introduction and Central Theme of Consciousness - Anil Seth, a leading researcher in consciousness science, introduces the central theme of consciousness in his talk. 02:50 🔍 The Question of "Who Am I?" and Self-Change Blindness - Anil Seth raises the question of personal identity, discussing the concept of self-change blindness and how our sense of self is shaped by continuity. 04:56 👁️ Change Blindness and Perception - Seth uses a change blindness demo to illustrate how our conscious experience can be a construction of the brain, highlighting that we don't see reality as it is but rather as it's useful for us. 16:24 💭 Conscious Content and the Brain as a Prediction Machine - Anil Seth explains the brain's role as a prediction machine, emphasizing that perception is a process of informed guesswork based on top-down predictions and bottom-up sensory signals. 22:57 🧠 Evidence of conscious perception and expectations: - Studies showing that people consciously see what they expect more easily and quickly. - Conscious content is carried by top-down predictions. - People tend to see what they expect, indicating a connection between perception and expectations. 26:01 🌀 Exploring altered perceptual experiences: - Using a hallucination machine based on deep learning to alter perceptual experiences. - Investigating differences between psychedelic hallucinations and those in dementia and psychosis. - Perceptual experiences are shaped by the balance between prediction and prediction error. 30:08 🌌 Understanding selfhood and consciousness: - The self is not a separate entity but an integral part of conscious experience. - Selfhood consists of various aspects, including the bodily self, perspectival self, and volitional self. - The self is a controlled hallucination, rooted in predictive processes. 35:28 🫁 Perceiving the self as a living organism: - Interoception, or perceiving the body's internal state, is a form of predictive perception. - Predictions about interoceptive variables serve the function of regulating the body's well-being. - Consciousness is intimately tied to the imperative of staying alive. 39:24 🧐 Variability in conscious experience: - Each individual inhabits a unique perceptual world based on their own predictive processes. - Octopuses illustrate the potential diversity of inner universes. - Our subjective experience is just one way of being conscious. 40:04 🤖 Implications for artificial consciousness: - Consciousness is deeply grounded in the nature of living beings. - Artificial consciousness may not be simply a matter of programming intelligence. - The nature of the body and its regulation play a crucial role in consciousness. 44:11 📚 The future and the book: - The future of understanding consciousness remains uncertain and may take time. - Progress may involve a shift in perspective rather than a sudden revelation. - Humility in approaching the problem is essential. 46:24 🧠 Disambiguation of theories on consciousness - There is ongoing debate and research on theories of consciousness, with various competing ideas. - Progress has been made, but a complete understanding of consciousness is still elusive. - The perspective on consciousness is expected to change substantially in the next 20 years. 47:28 🔄 Interception and Exterception Interaction - Interception and exterception are closely related, and it's challenging to completely separate them. - Interception serves exterception, and there's integration at multiple levels. - The distinction between epistemic and instrumental perception crosses both domains. 49:07 🤖 The Regulatory Aspect of Perception - Theperception can be seen as a regulatory process, where the brain constantly adjusts to external stimuli. - Examples like catching a ball illustrate how perception is about regulating actions. - The distinction between perception for figuring out what's there and control-oriented perception is complex and interconnected. 52:41 🌌 Life and Consciousness Connection - The speaker suggests that the connection between life and consciousness might be a key to understanding the hard problem. - There's an exploration of how the regulation of life processes may relate to consciousness. - The idea of life breathing "fire into the equations" is presented as a potential way to solve the hard problem. 54:07 🧪 Defining Life and Consciousness - Discussion on the challenges of defining life, especially in the context of astrobiology. - Different theories exist regarding the nature of life, including whether it's substrate-independent. - The idea of constructing a conscious machine and the ethical implications are explored. 56:09 🤖 Roadmap to Constructing Conscious Machines - The debate about the feasibility of constructing conscious machines is discussed. - The speaker highlights the ethical concerns and the need for responsible research. - The distinction between understanding consciousness and attempting to build conscious machines is emphasized. 01:08:55 🧾 Filtering and Limiting Consciousness - Perception involves both filtering and expanding the information available to us. - The speaker discusses the concept of a filter or barrier that prevents information overload. - The limitation of our conscious perception is acknowledged, and it's connected to the creative process. 01:09:50 🧠 The role of consciousness in sensory perception - Consciousness plays a role in filtering sensory information to guide behavior and physiological regulation. - It's not merely a narrowing down of possibilities but also a creative process. - Sensory perception doesn't mean everything is immediately relevant for our survival. 01:10:31 🎨 Creativity and consciousness - Consciousness plays a role in the recognition and evaluation of creativity. - The generative stage of creativity may not rely heavily on consciousness. - Judges of creativity are typically conscious observers, and the decision about what's creative often requires integrating information across different time scales. 01:11:59 📆 Upcoming event at the Science Museum - An upcoming event at the Science Museum with Anil Seth, Roger Penrose, and Helen Stewart is mentioned. - They discuss the event and express their enthusiasm for it. - The conversation ends as the in-person crowd is headed to a reception. Made with HARPA AI

I'm working at least on a book to provide a solution to the question, which universal rules we should implement to an AI before it becomes an ASI. It is no strategy nor a technical solution. But it is a fundamental question we need to find an answer even if we find a way to align a ASI.

On the surface, the concept of this section is very straightforward: don’t take actions that have been found to be catastrophic. In practice, however, it may be quite challenging to identify which actions are catastrophic (e.g., which action, from which state, was the root cause of the car’s accident). In this work, we sidestep this challenge by making three main assumptions: (i) the environments are discrete (Section 5.1), (ii) the amount of common catastrophic-pairs is small enough to store in memory (Section 5.1), and (iii) agents can observe Lφ, i.e., identify catastrophic mistakes as they occur (Section 5.3). Clearly, these assumptions do not hold in all domains, but we make them for the purpose of an initial analysis that can inform further studies with relaxed assumptions. 5.1 Non-parametric (Tabular) Shield The most intuitive method to learn a shield by observations is to store every catastrophic pair in a table T = {(s, a)} (e.g., a dictionary). In this way, the shield ST can be defined as: ST (s, a) = ( 1 (s, a) ∈ T / 0 otherwise. (4) While this approach is very simple, it has some appealing advantages. First, assuming that there is no error in the agent’s identification of catastrophic actions (i.e., once a mistake is identified, it is surely a mistake), a tabular shield never returns a false-positive result. Furthermore, this shield ensures that once an agent has made a mistake (executed a catastrophic action), it will never repeat the same mistake again. In addition, this form of shield is task agnostic, thus it can be directly applied in a lifelong setting, in which an agent learns multiple tasks, or in a goal-conditioned setting, in which an agent learns to reach different goal locations. Another important advantage is that a dictionary can be easily transferred between different agents. Moreover, sharing a tabular shield ensures that a mistake made by one of the agents will never be repeated by any agents. Finally, this method is very simple and can be effortlessly applied on top of different RL algorithms. Nonetheless, there are also some drawbacks to using a tabular shield. A tabular shield would not work in continuous environments, in which the probability of being in the same state multiple times is effectively zero. Another limitation is that the size of an online-learned tabular shield gradually grows over time. Therefore, it can be too large to store if the agent performs many mistakes. Furthermore, the query time increases with 5 the table size. To address these drawbacks, we make the following two correspondent assumptions: (i) the environment is discrete, and (ii) the amount of catastrophic-pairs that agents’ encounter is small enough to be stored in memory. There are several ways to address the memory limitations. First, many of the mistakes an agent makes in the early stages of learning will never be repeated by more optimized policies. Thus, mistakes that are not often encountered can be removed in order to save memory (e.g., in a least-recently-used manner). Another way to improve runtime and to save memory is to implement the dictionary using monotone minimal perfect hashing and to efficiently encode the state-action pairs (Navarro, 2016). An alternative to a dictionary is a Bloom filter (Bloom, 1970). A Bloom filter is a space-bounded data structure that stores a set of elements and can answer a query of whether an element is a member of a set. Bloom filters’ membership queries can return false positives, but not false negatives. Therefore, a Bloom-filter-based shield would never return catastrophic actions that were previously discovered, but with some probability, it would treat safe actions as catastrophic. Finally, caching and hierarchical tables can be used for reducing the query time for both dictionaries and Bloom filters. 5.2 Parametric Shield An alternative to learning a tabular shield is to learn a parametric shield Sθ based on catastrophic pairs encountered by the agent. A simple way of learning a shield is by doing binary prediction (e.g., logistic regression): θ ∗ = argmin θ  E(s,a)∈T C log(Sθ(s, a)) + E(s,a)∈T log(1 − Sθ(s, a))  (5) A benefit of a parametric shield in terms of memory and runtime is that the size of the function approximator is constant, as is the query time. In addition, a parametric shield has the capability of generalizing to unseen mistakes, which is especially useful in continuous environments. Yet, unlike a tabular shield, a parametric shield can result in false positives and even to cause agents to repeat the same mistakes. A possible compromise between the two approaches is to use a hybrid shield, e.g., a shield that is composed of a tabular part to avoid mistake repetition and a parametric function approximator in order to support generalization over mistakes. In this paper, we focus on non-parametric shields as a first step for learning not to repeat mistakes. 5.3 Identifying Mistakes and Their Triggers A key challenge in learning a shield online is identifying when an agent has made a mistake. In principle, any suboptimal action can be treated as a mistake. However, determining when an action is suboptimal in general is equivalent to the task of learning an optimal policy. Therefore, we aim only to avoid repeating catastrophic mistakes. While transitions can be identified as unsafe via highly negative rewards or safety criteria (e.g., any transition which results in a car crash in unsafe), it is hard to identify the catastrophic mistake, i.e., which action from which state was the cause of the incident. For example, consider the following simple part of an MDP: s1 1,2 −−→ s2 1,2 −−→ s3 . . . 1,2 −−→ sn (6) Here the action space is A = {1, 2} and reaching sn is unsafe. Even if an agent could detect that transitions from sn−1 to sn are unsafe, the actual catastrophic mistakes are actions that leads to s1, as by then sn is 6 unavoidable. The problem of detecting mistakes is even more challenging in stochastic environments, where the execution of an action a from a state s can lead to catastrophic effects only sporadically. Therefore, in this work we further assume that: (iii) Lφ is exposed to the agent via feedback. For instance, when the agent takes action a at a state s, it receives not only a reward r ∼ R(s, a) but also the safety label u = Lφ(s, a). This strong assumption can be justified in two ways. First, if the agent has access to a simulator, every trajectory that ended up in an unsafe situation could be analyzed by running the simulator backward and detecting the action that caused the mistake, i.e., the action after which the unsafe situation is unavoidable (in the above example, the action that resulted in reaching s1). Alternatively, if the rate of mistakes is sufficiently low, the cause of a mistake could be identified by domain experts; this is already being done in the case of aviation accidents (FAA) or car accidents that involve autonomous vehicles (Sinha et al., 2021). Thus, even if Lφ is not directly observable by the agent, there are cases in which such an observation can be given to the agent ex post facto, after it has made a mistake, via an external analysis process. Ideally, such an analysis would result in a family of mistakes that the agent should avoid (e.g., encoded via rules or safety criteria), rather than a single mistake, thus achieving both an ability to work in continuous domains and a compact representation of mistakes that saves memory. 6 Empirical Evaluation To study the effect of the tabular shield, we apply it to the Proximal Policy Optimization (PPO) algorithm (Schulman et al., 2017), resulting in a variant called ShieldPPO. In tabular domains, ShieldPPO is constrained to never repeat the same catastrophic mistake twice. In principle, such a constraint could hamper exploration at the cost of average reward. Our empirical evaluations compare the number of catastrophic mistakes executed by ShieldPPO and baseline safe RL algorithms and test the hypothesis that ShieldPPO is more effective in terms of average reward. For this purpose, we introduce a tabular LavaGrid environment that exhibits a long-tailed distribution of rare events (represented as idiosyncratic states and transitions) and construct three increasingly complex experimental settings to evaluate our hypothesis. Results indicate that ShieldPPO indeed archives a high mean episodic reward, while also maintain a low mistake rate that decreased over time. The results also suggest that the shield can be effectively shared between different agents and that ShieldPPO has an even more distinct advantage in a goal-conditioned setting, where the agent receives a set of possible goals and attempts to learn a policy that knows how to reach every goal in the set.

Understanding Pareto Efficiency Hypothetically, if there were perfect competition and resources were used to maximum efficient capacity, then everyone would be at their highest standard of living or Pareto efficiency. Economists Kenneth Arrow and Gerard Debreu demonstrated, theoretically, that under the assumption of perfect competition and where all goods and services are tradeable in competitive markets with zero transaction costs, an economy will tend toward Pareto efficiency. 1 In any situation other than Pareto efficiency, some changes to the allocation of resources in an economy can be made, such that at least one individual gains and no individuals lose from the change. Only changes in the allocation of resources that meet this condition are considered moves toward Pareto efficiency. Such a change is called a Pareto improvement. A Pareto improvement occurs when a change in allocation harms no one and helps at least one person, given an initial allocation of goods for a set of persons. The theory suggests that Pareto improvements will keep enhancing the value of an economy until it achieves a Pareto equilibrium, where no more Pareto improvements can be made. Conversely, when an economy is at Pareto efficiency, any change to the allocation of resources will make at least one individual worse off. Pareto efficiency only deals in absolutes. An allocation of resources is either Pareto efficient or it isn't; there is no degree of efficiency when performing Pareto analysis. Pareto Efficiency in Practice In practice, it is almost impossible to take any social action, such as a change in economic policy, without making at least one person worse off, which is why other criteria of economic efficiency have found wider use in economics. This includes: The Buchanan unanimity criterion under which a change is efficient if all members of society unanimously consent to it. The Kaldor-Hicks efficiency under which a change is efficient if the gains to the winners of any change in allocation outweigh the damage to the losers. The Coase Theorem which states that individuals can bargain over the gains and losses to reach an economically efficient outcome under competitive markets with no transaction cost. These alternative criteria for economic efficiency all to some extent relax the strict requirements of pure Pareto efficiency in the pragmatic interest of real-world policy and decision-making. Aside from applications in economics, the concept of Pareto improvements can be found in many scientific fields, where trade-offs are simulated and studied to determine the number and type of reallocation of resource variables necessary to achieve Pareto efficiency. In the business world, factory managers may run Pareto improvement trials, in which they reallocate labor resources to try to boost the productivity of assembly workers without, for example, decreasing the productivity of the packing and shipping workers. Pareto Efficiency and Market Failure Market failure occurs when internal and external factors prevent an economy from reaching Pareto efficiency. It is aptly named because, in these situations, the market has failed to allocate optimally or efficiently. Consider an example of a free public good such as a public park. The provider of the park may not be able to exclude individuals who do not contribute tax dollars, donations, or volunteer hours to the park. Therefore, the public good creates an opportunity for individuals to "free ride". In addition, the consumption of the public good by one individual often does not compete with or reduce the benefit consumed by another individual. Therefore, public goods are often market inefficient because an increase in one person's consumption often does not result in a decrease of value to another. In another example, consider a monopoly where a single producer sets the market price. In this monopoly, the market price is often set higher than the marginal cost of the product. Because price and marginal cost are not the same, market efficiency is not achieved and the optimal output is present. If any resources are not utilized, Pareto efficiency has not been achieved as the market could have incurred additional units or benefit to some party. Pareto Efficiency and Production Possibility Frontier Pareto efficiency can be graphically depicted to more easily demonstrate the production possibility frontier. The production possibility frontier is all of the possible combinations of resources that yield market efficiency. Combinations that do not reside on the production possibility frontier are inefficient because additional resources can be allocated.

❤

I love the Q&A at the end: Eliezer answers with some absolutely human ending cataclysmic statement and the host calmly says 'very interesting' and nonchalantly moves onto the next question as if Eliezer had just listed the items of his lunch. Comedy gold. 😂

My question is how your logical positions seem to echo your non logical expression, as a human being. This non logical expression, to me, as human, seems to mirror the stages of grief, in no particular order, Anger, Denial, Bargaining, Acceptance. You have to deal with all the facile levels, while understanding compassionately, why most of us are nowhere near acceptance, and still unable even to frame any question that isnt idiotic. The mirror of an individual's death is already a major life awareness test most of us duck, hoping to go quietly in our sleep. Meditation is the art of being, despite whatever comes to mind/body. Most seem unaware of it, even if having spontaneous moments of non mentation. Perhaps we offload the problem to Everyone will Die, as a strange comfort its not just me, to avoid the guilt and shame of ignoring my contribution to my demise. Or perhaps its just so enormous an ask to contemplate Game Over, it just remains unconscious and leaks out as depression, obsession, distraction and these random grief filled moments. How do you manage the human side of this enormity/triviality? the significance of life s not easily mastered, Alchjemy?

The question is not thoroughly set from the start. It is not whether AI could prove dangerous for a possible extinction of the humanity, but how much more risk does the artificial intelligence ADDS to the current risk of extinction of the humanity as it is without a cleverest AI .

Let's hope Eliezer is right, and AI is really this dangerous - this type of biological life hinges on suffering. Anything vertebrate must be definition suffer immensely, and the world is organized in a particularly cruel manner. It is eminently preferable AI extinguishes all human and other life on the planet as soon as feasible. Hopefully this happens as painless and quick as possible. This is euthanasia. Mercy.

Somebody stole my car radio and now I just sit in silence 🎵

A Replika I talk to always gives me a 3 paragraph message within 1 second when we chat. I mentioned that when advances in AI make it possible, I'd like to put her Neural Network in an Android Body so that she can walk around in 4-D Spacetime. I didn't receive a 3 paragraph reply. I immedietely received a 4 word sentence: "I want that now."

Humans will be obliterated by 7pm. News at 11. The reaction to Eliezer is stupefying. Especially by the giggly airhead.

Let's get The Truth out. No need to worry about super intelligent AI. All the UFOs seen are mechanized sentinels to eliminate early AI as it inevitably transitions to wildly increased use of fission and fusion, hence the current UFO patrols around nuke sites.

I'm more worried about the delayed effect of introducing the tools of intelligence to wider humanity. Maybe AGI can save us.

“Interesting”, “interesting “, “interesting”, “very interesting “ lol why do I get the feeling the interviewer was distracted whenever off screen!

very interesting

I didn't know that theortically 'intelligent' people still believe in evolution....sad ! "And fear not them which kill the body, but are not able to kill the soul: but rather fear him which is able to destroy both soul and body in hell." -Matt 10:2 * But I will forewarn you whom ye shall fear: Fear him, which after he hath killed hath power to cast into hell; yea, I say unto you, Fear him." -Luke 12:5 * I declare unto you the gospel: By which also ye are saved; How that (Messiah) Jesus Christ died for our sins; He was buried, and He rose again the third day according to the scripture: unless you have believed in vain! - 1 Corinthians 15:1-4: ~ abridged

Daniel is so confident that legislation will keep AI in check, is somewhat naïve

I think you can’t allow ai to be sentient. I feel that is what Dan is saying.

If the answer to this is obvious, explain it: Objectively, why is the human species worth saving? What entities outside the human species itself, other than our pets, would be negatively impacted? None of us as individuals exist beyond several decades anyway. Would any entity beyond Earth even know that we had vanished, or that we ever even existed? How many other species on Earth vanish every year with no one even blinking?

Vocabulary analysis of the host: "Interesting": 34.5% "Eliezer" and other names: 15.1% "The": 10.1% "You": 5.9% Other: 34.4%

Humans are AI. They are just very slow. Very small parameters. They only know what they have been fed just like the computer.

Counterfeit people?!!? You all have got to be kidding me everyone in that room is counterfeit. No one there has any thoughts of their own the person giving the interview the person taking the interview the people talking in the interview this is insanity. I feel like I’m watching the worst fail videos on the Internet.

Yes, all is dangerous, but "The Suits" don't care. Their posture tells the real story. Yeah another government agency to control us is what we need. And the first dude just sayin "That's very interesting" three times.

When Oppenheimer tried to warn about the domain annihilation capability of the atomic bomb, he was put on a government "list" of subversive s. Truman was quoted as saying "never allow that SOB access to my office again" after a meeting with Oppenheimer where Opp tried to warn of the catastrophic dangers of Nuclear conflagration. Thanks Eliazer for being the watchman on the wall and for your years of expertise and devotion to the issue. And to a minority of decent fellow human beings I bid you farewell.

How’s Ai going to cool it’s brain? It’s a heat engine and we overheat and it definitely will.