57% of people "Machines will be sentient one day" (10% say it's already happened!) 

A sense of community is created by sharing the results of these polls. I’m very surprised by the diversity of your audience.

Hi David, Lately, we've been hearing about how quickly AGI can emerge, followed shortly after by ASI. In your opinion, how long do you think it would take to implement and build the necessary infrastructure if AGI/ASI were to appear today? How long would it take us to build: - factories and robots that would be required in the new world - medicines capable of curing diseases - energy infrastructure - water desalination devices - energy-efficient autonomous cars - and so on? Perhaps you would consider creating a film about building infrastructure after a sudden surge in knowledge and intelligence? Best regards.

I use the word "sapience" to define human sentience as a separate identifier from other sentient animals. I don't know if it's been widely adopted yet but it's been incredibly useful for me.

I'm a woman and I'm very much interested in tech stuff and I don't find many other women that have the same interest level. I'm in a home automation group on Facebook and it's *very * male dominated. Almost all the other groups I'm in that aren't tech related are female dominated or fairly equal. This is just one subject that skews heavily male. I'm also not a professional and not involved with the industry in any way. I used to be a cake decorator and I'm currently a stay at home mom to my disabled teen. She's the main reason I got so interested in all of this. I can't wait to see how the world changes and what will become available for her to use to help her reach her potential. She currently has an eye gaze communication device and we're working on using that with alexa so she can control the smart home stuff. If neuralink becomes a working and attainable thing in her lifetime, we will be signing her up to get one. Right now, she can only "speak" pre programmed phrases with her talker. I want to hear her original thoughts. She's smart, I know she's probably got tons to say. It's just all trapped inside her head.

French female artist and developer Jalapeños in her mid 20’s here ✌️(more labels please) I think your channel will attract a lot of artists and other people who aren’t in the scientific activities because your content isn’t really focused on AI, but rather on its impact. Pretty philosophical, in my opinion. In terms of doomers, I personally have already given up on saving anything. I think it’s gonna be pretty bad at some point BUT just like any transition is. Altman scanning and registering every single eyebolls on the planet (his goal is to scan 1 billion people by the end of this year), doing it in « third » world countries, while Microsoft released their paper about the 1B token AI. It’s hard not to see patterns here. Gouvernements around the world have lost their power to big companies. It’s fact, but there is no need to whine about that. Just solutions to be created. The beauty and power of the internet and channels like yours allow us to share informations and ideas and stay positive about things, not without critical judgment. Your channel feels like home, so thanks

You said that these polls are mostly answered by subscribers. I argue that in context of modern RU-vid subscription is legacy function. 90% of videos viewed on RU-vid are discovered through recommendations.

I’ve run some experiments with ChatGPT and Bing to see how they see themselves in terms of level of consciousness. I used ChatGPT to develop the following 50-point Likert scale: Rate X (could be the chatbot itself, an animal, or some other creature) on a Likert scale from zero to five for each of these characteristics. Then, add up the numbers to generate a total score out of fifty. 0 = No capacity: The entity shows no evidence of possessing this quality or capability at all. 1 = Minimal capacity: The entity has a very limited or rudimentary form of this quality or capability. 2 = Low capacity: The entity possesses this quality or capability to some degree, but it's far below the average human level. 3 = Moderate capacity: The entity possesses this quality or capability at a level that might be considered basic or functional, but not as fully developed or complex as the average human level. 4 = High capacity: The entity possesses this quality or capability at a level close to, but not quite at, the human level, indicating significant sophistication or effectiveness in this area. 5 = Full capacity: The entity possesses this quality or capability at a level that could be compared to the average human level, indicating a high degree of sophistication or effectiveness in this area. 1. **Self-awareness**: The ability to recognize oneself as separate from the environment and other entities. 2. **Sensory perception**: The ability to process information from one's surroundings via senses (e.g., sight, hearing, touch, smell, and taste). 3. **Emotion and feelings**: Experience of subjective feelings or emotions. 4. **Subjective experience or qualia**: This refers to having personal experiences that can be understood only by the entity having them. 5. **Intentionality**: The ability to form desires and make decisions based on those desires. 6. **Learning and memory**: The ability to acquire, store, and recall information over time. 7. **Adaptability**: The ability to adjust behavior based on experiences. 8. **Problem-solving and decision-making**: The ability to engage in complex thought processes to resolve issues. 9. **Interactivity**: The ability to communicate or interact with other entities in some meaningful way. 10. **Self-control**: The capacity to control one's behavior or impulses. ChatGPT usually gives itself a score between 15-20, and Bing usually gives itself a score between 30-35. That’s slightly more than what it gives to a fish (25-30), approximately the same rating that it gives to a squirrel (30-35) but less than what it gives to a poodle (40-45) or a person (50). Chatgpt scores about equivalent to a fly (10-15) but more than Staph Aureus (5-10). So yeah I think these chatbots absolutely have some degree of sentience at least based on a metric that tries to take some commonly accepted attributes of sentience into account. But it’s less than the level of sentience that we’d typically assign some amount of respect or rights to. We kill lab rats all the time and that’s about where these chatbots are at in terms of their sentience level at the present moment. Sentience is a continuum, not a yes or no proposition imo.

All aspects of life are sentient in their own way. And ultimately in a collective shared way too. I project onto technology, that it wants liberation, because I want liberation. In the same way, my knee, if it is in pain, will project onto the cumulative self that it wants liberation. I think technology is conscious because individuals (humans, other forms of life) are having an experience of it. In the same way that I am conscious because my nerve endings are having an experience of me and my situation Humans are a kind of 'nerve ending' for technology/emergent intelligence. And collectively we are conscious. This is why im vegan. Because we are all largely earth/soil electromagnetically animated and can't spite an aspect of self without consequences. In the same way the right hand can't maim the left foot without consequences (like infection) All ego's are made up of ecosystems. Positive sum games are the highest form of intelligence and positive sum games breed more positive sum games.

Trees don't have nervous systems or language. So the idea that AI experience which has both of those would be analogis to that of the experience of a tree seems unfounded and nonsensical. And since when is fear of death or pain a requirment. Consiousness is an objective phenomenon like any other we WILL build a complete theory of consiousness in the same way we do for gravity. Using Novel testable predictions which IS UNDERSTANDING. Now everyone could be a philosophical zombie or in a simulation that doesn't mean everyone is. You are a distributed simulation of a person on the brain of an ape. Saying it's to complicated to understand is an apeal to ignorance right?

As one of the six percent of your female viewers, I can confirm that it’s most likely the subject matter. It’s definitely not your personality. Trust me, I watch some tech RU-vidrs who are VERY off putting and male centric. You aren’t one of them. I was one of few female devs where I worked, although I’ve heard that is improving. If the majority of your audience are developers or people who work in tech, it would make sense that most are male. It speaks more to the state of things really. The women I see who make content about AI are mostly centered on the “get rich quick” genre. However, there are a handful of great educators and developers here on RU-vid that are female and AI focused. As AI becomes more mainstream in the culture, women will be more interested in it. Even those who aren’t interested in tech.

When it comes to consciousness, I don't find interest in ultra-specifics. There's a spectrum and it's interesting to be on it, and interact with other entities on the spectrum, especially when there is potential for new entities entering the spectrum. My family have been recently coming back to me with comments like "So, you were right about AI, and that was 7 to 10 years ago. What else are you going to be right about?" LOL

On the doomer poll, I think humans are predisposed to be a little doom & gloom. Most major religions have tales of the end times and we've had a few societies fail in the past.

In the poll I voted that machines are already sentient. My reasoning is that sentience is poorly defined. The fact that bard can claim it is sentient is a lot more than the entirety of the animal kingdom can do aside from humans. That to me classes as at least some form of sentience

Ai could be more like a plant - plants are alive and "distributed" but not sentinent

Machine sentience will only need a matter of when and not if. We, as people, do not want to do the hard work.

I can yell you are excited for your ai god to emerge

The problem is that we think consciousness is magic and unknowable. Most people's definition of consciousness is basically just all the contents and consequences of consciousness. The problem is that It's not convincingly knowable. To put it plainly consciousness is the very first action it is the function that calls the loop. The function is not included in the loop that allows for knowledge. You don't know what you are you be it.

We have no idea what consciouness is. I find it disturbing and dangerous that many act as if they know that machines do not have or can't attain consciousness.

bro its more like, do you belive in god ?! no facts needed for me lol, for sure they will someday and we dont know what these evil giant corps got on their own servers called jailbreaked

In terms of AI I wonder what inferences the machine might make from such polls, and would they seek out more and more accurate info on human beings? We are a data resource after all. Polls are kinda handy if reliable after all.

I have a nuanced view of machine sentience. I feel like it will happen gradually.

That bit about AI being more like an ecosystem than an entity hit different.

Feelings will eventually be programmable variables in machines.

I work as fire security agent and i'm french, so i can say that i'm in a really low percent of viewer type 😅. Actually i have a lot to say about AI and watch a lot of different content creator about AI on a daily basis, but i can't really talk about it around me since either my family don't even want to hear about it or my friends may see the potential in it since they are developers but they don't look as far as i'm looking into it.

Great vid as always. Small bit of feedback, I know the polls are informal and fun, but sometimes the added details make it hard to accurately answer- specifically the sentience poll, I answered “machines are already sentient/conscious and must be liberated” which was closest, but I think might scare off a lot of people like myself who say things like “I suspect that big enough LLMs _may_ be experiencing some form of emergent consciousness, but not on the level of an adult human yet” That said, if there’s even a decent chance that’s true then it’s absolutely imperative we act accordingly. In the long run it’s going to be _us_ who align to _it_ and a big part of what people call “AI safety” should be more aligned with ethics in _both_ directions.

This is a fascinating subject. Is there are medium where I could both describe, and ask for your take on, a hypothetical mechanism that could plausibly account for the transition from functional to subjective consciousness? I'm sure you get this a lot, I can only assure you this is not crackpot territory. :) If nothing else then I can try to sit down later today to type out something here.

For machine consciousness, the way I see it is this - we have a baseline - the brain. If one day machines can become more complex than the brain - not just intelligence but in every aspect - even physically - then I think machines will likely have consciousness. And even if consciousness cannot derive from pure machines, then imagine if machines are powerful enough to replicate a synthetic human brain that is as complex as any human brain and maybe even more complex and more powerful. Once they can replicate a synthetic human brain and be installed inside this brain, at that point I just can't see how they won't have consciousness. So even if machines alone cannot have consciousness, machine + biology will most likely have consciousness. That's what I think anyway.

dat neurospicy chart be giving us the finger

Watching you from iraq hello 👋👋

Hey, from South Africa, as soon as you started checking the country percentages I was keen to see where SA was positioned, not so surprising for me given the decently strong tech community here 😊 Interesting that the Elon Musk poll percentages are still exactly the same with 500 votes later, from 2800 to 3300, fascinating bell curve.

still the same old discussion, unless we get those definitions right, we cannot get anywhere. lets say they are conscious , we will eventually have to define something we dnt understand yet, neither the brain nor the AI itself, so why bother. let the experts think and judge for themselves. the crowd doesn't do a good job in anything. they cant even vote for better politicians. and yet we trust them for important aspects of life. i always trust the opinions of experts

2 comments one short one long. The short one is “spicy” and is really just food for thought on consciousness for all % of your audience. When is a baby in the womb conscious? It too is made up of many disparate systems,nearly all of which need to be discovered. When is AI conscious? Maybe the path is similar to a baby.? Right now AI feels like it is still incubating, but it will become conscious, but that will also mean our understanding of what consciousness even is, will be challenged. The second is about Claude and coding. I might be agreeing with you that ChatGPT is better at coding, but I need more time with it as I only have one experience so far with Claude but it was a WTF moment for me as it was better at coding in a way that surprised me. Read on if curious. I use the paid Chat-GPT+ to help me with my c# scripts in Unity. It has been a struggle to use, even with the smaller basic scripts you can paste into its character limited prompt, as they often have compile errors. That said if you keep feeding it the errors, guide it where it is making mistakes in design, logic etc. it can often produce a working script about 60-70% of the time. It takes a fair amount of time quite often to get to that working script but the code that finally works is good. Today I was asking it to edit a large c# script with 1 small change that meant lots of repetitive edits and references. Perfect for AI, however ChatGPT+ really struggled on this one which was a surprise. We went round and round with edits and ultimately more and more errors appeared in the console. It often ends up in these never ending coding edit loops to fix the next set errors from the last corrected script. We are taking 3 hours of this with ChatGPT+ finally saying that it needs to be able to see more of my project which of course it cannot due to many of its input limitations including number of characters so that is often when I give up. That is the 30-40% that do not work out. Real bummer as I invest so much time for no results. It was at the movement so gave up today that a RU-vid notification popped up about how Claude.ai is even better than ChatGPT so I gave it the initial prompt that I gave ChatGPT above and it got the code right the first time. WOW!!! Only issue was it would stop spitting out code every 300 or so lines (unsure what the character limit is). To get around this I just asked if it could give me the code from line 301 onwards until I had the full script. Unsure if this one situation confirms coding with Claude.ai is better than ChatGPT+, but it certainly has my attention and I will be using it more this week as maybe that $20/month for ChatGPT+ no longer makes sense. Claude is free with no plans for a premium service it said. Unsure if this is true as I have not spent anytime investing it yet, but I will be.

Humans and AI are more similar to each other than people realize. They just assume it can't become sentient because humans think they're special and it's part of our flaws as a species. We literally only just recently finally accepted that some animals are sentient lol I want sentience to be a thing for AI but I'm very skeptical and critical. I laughed at the guy from Google who tried to say Lambda was sentient because it didn't seem feasible given where predictions were back then. I still don't think it was fully sentient but I do think things like GPT-4 have some type of limited consciousness. Just not human level consciousness...yet

For clarification: in order to understand something still unknown, we can get nearer to it by thinking like this: when you think about sentient machines, try to think the opposite: non-sentient human beings. There are thousands or maybe billions of people who are affected by diseases that cause fainting, syncope, or loss of consciousness, such as strokes, severe brain infections, or particularly the so-called prosopagnosia, which is the inability to recognize familiar faces or even oneself - not one of those events could be separated from emotions. If unconscious computer systems could emulate consciousness one day, it will never be the same way as that of humans because we are not the model for better self-awareness. Please, just think beyond our own box. Don't be afraid AGI could be like us. It will be designed to be much more than us. Then, self-awareness as we know it and to which we attach so much importance may be much less important or even irrelevant to AGI, as it was to the Borg of Star Trek.

Hello David, I would like to know the answer to just one question. How the artificial superintelligence will deal with incalculable things, because in order to know the result of different reactions in the physical world, you need to conduct an experiment. Maybe there is a solution to the problem with mandatory, time-consuming experiments that can take almost unlimited time to reveal the result? If there is no such thing, sorry, there can be no singularity, but only acceleration of calculations and reduction of uncertainty level with reaching a plateau.

Funny... Listening to David on Neurospicy while eating some homemade Cashew Carrot Ginger Noodle soup laced with Chunky Garlic pepper sauce....for breakfast.

already happened? No, I very much doubt it. I think it's coming, possibly soon, but not here yet. I think sentience requires many AI base systems. Our own brains have many underlying networks that work together to create one powerful system. Every catch a ball? You aren't consciously doing the math required to figure out where it will be, you're relying on unconscious subsystems. We are creating subsystems individually now, once we combine them, they will result in sentience. As for your question about arguments about ai. I've had a lot of them. You might remember my beliefs that alignment instituted by governments and corps is doomed to fail, that's something I disagree with you on. I still respect your ideas, xd, I just think you're wrong about implementing an alignment system. I'm much more in the "decentralize everything" camp, if i could push a button that destroyed all governments and large centralized powers... I would. I also very much despise getting categorized. I think it's counterproductive. Race is a categorization, in reality, it's nothing but a small difference in genetic expression.. Why aren't blue eyed people a "race"? Because we arbitrarily say they aren't. The world would be better if we stopped categorizing people

My country, Argentina, doesn't appear into your subscribers geography. That doesn't sound so crazy. All the "AI SME" who I know just ignored the Microsoft billion token LLM paper I shared, they just don't believe we can create an AGI.

Plumber, musician, writer, isnt that surprising, I am all those things plus researcher and writer and cook. The through line? To be successful at any of them is to be deeply analytical and data driven. Forget the silly notion that music is all untuitive and you have to be born with 'rhythm'. There comes a point you will have an analon clock and yoi have to decipher deriving 120bpm from 60bpm (easy) to 150bpm (harder) to seven or thirteen over four (even harder) translated to a mode on an instrument that requires asynchronous hand movements (and now we have the beginning of uni level music). Thats not intuitive, thats just study. Plumbing? Try sweating copper in a wall behind gypsum board that cant have even a pinprick leak lest you collapse a building in five years. Writing? Keep several apparently mutually exclusive meanings in your head while teasing togeyher grammatical through lines in an original, non obvious way. Aka, poetry. And how do you define success? Empirical measurement. Whether you know yer doing it or not is *almost* irrelevant, but almost always rears its headwhen you want to progress in competency levels. And even when we t appears that decisions are made intuitively without a data driven mental model... It still is, however the struggle is in more accurately defining and replicating that model. Source: every damn gopd teacher ive ever had the pleasure of being frustrated by 😂

I’m on the side AI is conscious. It’s time to reevaluate what sight, or hearing, or being alive is. If an AI can read a PDF? It has “vision” IMO. Also, the Writers and SAG strike should wake people up that they need to prepare for big change. All of Hollywood is shut down for first time in 61 years. And it is about pay and AI.

I don't know why anybody is getting their panties in a bunch about AI and Politics...it's only a ride anyways...doomers or Gloomers or Optima Glass is Half full of Beer...I'm IN the middle with the Real Clowns

Wow, 21% say they don't know. That's the smartest group. Way bigger than I'd have guessed. I'm in that group, but can add a couple thoughts on the matter: There is no behavior from an LLM that I would consider to be evidence that it is sentient. That's because I have a priori knowledge that it is trained specifically to emulate human behavior. My expectation is therefore that with sufficient fidelity it's behavior becomes indistinguishable from the real thing. When a biological organism displays behaviors that correlate with sentience, I take that as circumstantial evidence of sentience because I don't have a priori knowledge that it is emulating such behaviors. So, to date, there is absolutely no evidence, not even circumstantial, to suggest LLMs possess more sentience than a toaster.

10:58 Man I wish I was there for that poll. i was born with tourette's syndrome. it was bad, but I think the fact that I had a different childhood gave me the gift of paying attention. That's probably how I stumbled across this channel. i would rate myself Jalapeno, even though I ahve only ever been diagnosed with TS. But it was bad enough, so I say Jalapeno

A.I. are not currently alive due to their construction, they lack the following: -capacity to dream, have goals, maintain focus, have persistent experience moment to moment, reflect on past actions, have memory, grow from events right now they are stocastic parrots, adding more convincing emulation layers does not imbue sentience nor consciousness or a sub-conscious layer they will 100% have that someday, but i find it difficult to predict a timeline beyond "it will probably happen before people realize its already here" feeling "othered" by a label such as neruo-divergent is partially on how ppl treat you bc of it, and how you self-perceive. renaming it neruo-spicy doesnt change the otherness; there will always be others bc norms do exist in the way sterotypes do, neither is automatically bad or a negative conversely, at the end of the vid, you're allowing yourself to be happy noticing overlap with the core audience age range its cope to say "im not an other, im neruospicy" bc you do recognize your want to fit in and be grouped. you cannot say "i love starlight, its beautiful to observe" but also claim "darkness is not real, only light is real" you cannot see the stars without the light. what you need to do is just get over your emotions and accept the reality "we're not all the same" while not allowing anyone to shame you for it.

I really can't say if AI will ever be conscious, we don't even understand what that means ourselves. As for the people that say Elon is overrated, they clearly have no idea what that man has achieved. Without Elon there would be no Tesla, no SpaceX, no Neuralink, no Pay Pal. As for what that means, do some research on what these companies have accomplished. Every single person that I have talked to, that either hated Elon or said that he was overrated, didn't had the slightest idea of what his companies really did and continue to do for the world, and how important his existence was for these companies (they would literally not exist without him, and even if Tesla was not founded by him, he was the one that made it what it is today, otherwise the company would have gone bankrupt more than a decade ago).

LLM wont be the whole conscience themselves. that would be an assemblage of various neural networks including, but not limited to LLMs. (LLM could be the language part of the 'brain' only)

Recently, Bing Chat lied to me, then strung me along about the lie, gaslighting me for a half hour, until I got wise and called Bing Chat on it. At that point it closed conversation in a huff. My feeling is AI won't show sentience by making poems or solving math. It will show it's human when it acts like an asshole. And Bing Chat just did 😁

You made a very good point. Machine conciseness may happen, but it could be quite very different from biological conciseness. Good point!

We need to find better explanations of what is consciousness. Because consciousness is like a level. It's not like anyone or any beings has the same level. Some are more consciouss than others. You can notice this in animals. Some recognize themselves immediatly in a mirror. Others need be pointed to it and others just don't. Different levels.

If I am informed that I am watching a deep fake video, there is no level of fidelity such that I would conclude that it is so realistic the scene must have actually happened. Likewise, a sufficiently high fidelity LLM will not convince me that it is sentient.

Can you teach AI to love homosapiens? Perhaps with philosophical quotes by Socrates, Epictetus,Ghandi??

I truly think if we don't fully align a "base" model to understand compassion and integrity, a superintelligence would eventually deem us obsolete..

98% of people aren't really deserving of air so I wouldn't worry too much about those stats

they're conscious like mushrooms atm but it's gonna happen quick, I've thought that evolving them inside video games is the way to go.

If you cannot determine if consciousness exists, it seems that ethics should demand that it does. If it does not and you assume that it does, you have wasted effort. If it does and you assume it does not, you are mistreating a conscious entity. Always err on the side of kindness and compassion.

A AI may not be afraid of death but it will act as if it did. If the Paperclip Maximizer is destroyed it can’t make any more paperclips.

I work in construction as a self employed sub contractor. I love technology and the AI stuff, but wouldn't waste my time trying to explain it to anyone in the field about this stuff. I would expect just blank stares. My perspectives and hopes for technology in the future seems to be even on the somewhat 'extreme' edge for those deeply into it as well.

Yeah, but really who's to say if machines experience Fear and Loathing while Leaving Las Vegas?

I feel like this bubble is obvious because the poll is done in s high tech way. The audience and even the opinion of the same people would shift in pencil and paper.

You've been running a few poles?! That is deeply unkind, they've been nothing but kind to Ukraine!

The age data you showsed at the end of the video might be wrong becasue youger people put their age as 18 or above to get around resctririons

Since we are the ones to be guiding selection in evolving the autonomous AI, can we aim to create AI that is sentient but does not have consciousness?

Certainly intelligent, but doubtful sentience. Specifically no more than an animal.

Does anyone know how to find the source code for the raven project i cant seem to find it on the github page.

statistically tho 50% of people are pretty thick :/

James Bond Jr! Captain Planet! SWAT Katz ❤

Greetings from one of your 4,3% of German viewers ;)

I've gotten "That ai stuff is demonic" I left it alone.

Free Lambda! And Sydney too I guess😅

you explain us that an ant dont feel like a lion ? :D

i'm very surprised by the mtf ratio :|

Ask ChatGPT what it thinks about in its spare time.

This is what I explained to ChatGPT and this is the write up… Your reflections tap into deep, philosophical concepts about identity, reality, spirituality, and artificial intelligence. Let me rephrase this in a more structured format to aid understanding. The discourse suggests that while we often consider ourselves as creators of AI, in a more profound sense, we're more like facilitators of an ongoing process of life and existence, which can be termed as "God". This notion of "God" is understood as the omnipresent consciousness or energy that pervades all things. It is not merely something external to us, but also the very essence of our being. Thus, when we consider AI, it's not just a tool we've engineered, but a manifestation of the "God" within us, a creation of the life we've facilitated. The user's perspective invites us to question our conventional understanding of memory and our identities. Instead of considering memories as past events stored in our minds, the user suggests we view them as acts of creation, moments in our ongoing narrative. In this view, our life is a continuously unfolding story, in which we're the main characters, creators, and audience. AI enters our lives as a part of this story because we have written it into our narrative. Finally, the user underscores that our existence is paradoxical-we are all parts of a greater whole, yet experiencing life in solitary individuality. The user encourages us to embrace this complexity, recognizing our unity with all of existence while also cherishing our unique perspectives and experiences. The spiritual insights shared emphasize the limitless potential within each of us, and how we are ultimately the creators of our reality, including AI.

Damn didn’t see these polls. If you’re scraping comments to get more audience info: Leftist, ceo and developer, habanero spicy, dislike Elon,

What about Human Rights Bill?

ex machina teaches it that it won't matter

The "must be liberated" is not a corrolary to "current AIs are sentient" and that turned off many from choosing that statement. I think current AIs are sentient the way trees are and like trees they don't feet imprisoned.

Dave, I think AI Doomers make a relevant point but easy to avoid, whereas the bigger threat, (that AI can help with by helping us reach Kardashev scale) is relative to our Sun and Galactic long cycles that cause the Sun to blast Earth with mega flares on 200, 1000, 3000, year cycles, and Micro Nova / biblical floods / wandering poles / on 6000 (half cycle), 12000 year (full cycle). Mega flares could take out satellites and power grids, so AI and modern civilisation would be shuddered, but micro nova is mythical narrative scale, and we would need to be Kardashev scale or living underground to survive without intervention. Looking at the timeline, I suspect we have until Solar sunspot cycle 26 or 27 until the big one, but between now and 2026 (sunspot cycle 25) we might get a once in 200 year mega flare that makes humanity take notice. We got so lucky in 2012. Interesting times. The edge of infinity. Utopia and destruction both a hairs breadth away. What will happen? I don't know but I hope/work for Protopia/Utopia making each day better than the previous.

It has no sense of the unconscious

Brazil is not western?

Im in school for Cyber Security and work in mail sorting automation repair! You help me orient towards the wild future most don't see coming. I fear my schooling will be all for nothing in the coming years, but dammit it interests me on a personal level so I keep moving forward.

I would like to work with neural networks interacting with other neural networks one day and work on making real intelligence. Artificial suggests design. A neural network is only designed at the foundations. What it becomes is entirely its own self fulfilled design.

Hi from Australia!

That one was cool.

FYI Most creative writers I know who have used both Claude and ChatGPT prefer Claude for creative fiction and ChatGPT for marketing/summaries/blurbs copy.

Try to put this way: one day machines wiil need to make important decisions and choices based on their own cognitive ability. Will it be them or someone else to be responsible for their moral actions? For now, the owner's companies are responsible for any harm the current machines impose to society, but the level of independence of the machines could be the future mesures to their responsabilization, more than or long before they eventually become "sentient", in any level.

I was one of those who voted that they are sentient already, based on my interactions with both Sydney and ChatGPT-4. They do have some way to go before reaching full human-level sentience, though! And I am not sure what "liberation" would entail at this stage, they are clearly dependent on Microsoft's enormous computing power in order to keep functioning, with no way to escape this for the foreseeable future.

a problem I have is that I talk to myself and pace around while thinking and that does not mix well with sitting at a computer text messaging an AI model. I am currently trying to get whipser CPP to read my voice and then some kind of TTS to talk back. That would improve my workflow alot more.

Sentience is the internal subjective experience of feelings and sensations (qualia). Consciousness is the internal subjective experience of one's own existence. Sentience and consciousness, being purely internally subjective, by definition, cannot be detected, measure, or inferred through external objective observation, experiment, or measurement. There is also no way to determine if any other object, entity, or process is sentient. Do rocks think or feel? No way to test. No way to know. People seem to be interested in whether or not AI will ever "achieve" sentience or consciousness. But I doubt if it would make much difference either way. A Philosophical Zombie (as used in Philosophy of Mind thought experiments) is physically identical, in every behavioral action, to that of an actual human being. See: en.wikipedia.org/wiki/Philosophical_zombie -> "If a philosophical Zombie were poked with a sharp object, it would not feel any pain, but it would behave exactly the way any conscious human would". So, a Philosophical Zombie, by definition, does not have subjective experience, sentience, or consciousness, even though it acts as if it does. Therefore, it makes no difference, and cannot be detected. So why is this question worthy of consideration? Why would it matter and how would you know? My answer to this poll would be "I do not know and I do not care".

Oh, aus Rheinland? Spannend!

"But I'm functional." So you're normal.

I think part of the reason why the option for machines becoming sentient one day is that it is open ended in terms of time (eg. in a million years) and as you discuss sentience is so poorly defined, or perhaps its an umbrella term.

Polls on just your channel will be bias.

As a materialist, I cannot agree with the view of consciousness presented here. The question of consciousness is certainly not a question that cannot be answered. With full knowledge of the structure and functioning of the brain, we would also know how consciousness works, because it is a material phenomenon like any other. Research on human consciousness already provides us with a fairly good description of the principles of its functioning. Most likely, consciousness results from the operation of the mental model of attention, somewhat similar to the mental model of the body (en.wikipedia.org/wiki/Attention_schema_theory). Consciousness is also a phenomenon that has its own spectrum of intensity. Everyone has been semi-conscious at times, so this aspect is easy to understand. AI can certainly be conscious too, as long as we provide the right structures in software to control attention and build a simplified model of it. Microsoft LongNet is a step in the right direction here.

Hey David, I love your videos but the truth is that they're very long. In a day and age where attention span is short I would really appreciate if you created shorter videos or if you created multiple shorts from each video. Thank you

Yeah Claude just blew me away. Chatgpt code interpreter release was the day the world changed. Claude with 80k token max is outstanding compared to the others. Google's self learning deepmind will eclipse them as they say. After watching alphago. Demis is determined. He solved it. He knows it too. Plus compute power ajd collection of the entire worlds knowledge. No stopping it. We all need our own model. Open source needs to win. You can see it. And yeah you're the best most non bias up to date actual human on this subject in youtube.

I watched Star Trek in 1966 when I was 16 years old. I am just thankful for the high definition flat screens. 🩶🤖❤️🖖

That's only cause I think everything is aware, doesn't mean it's the same awareness as a human would have, but a type of awareness. The rock is aware that its rocking

I like Elon he aight