@@dylan_curious Only if you assume that agents will classify as a new species. I think 2026 might be a bit early, more like it will probably be undeniable by 2028
This is kind of how I experience the "AI isn't really intelligent" argument. Sorry if it's a bit reductive. AI SKEPTIC: AI can't be real intelligence. ME: Why not? AI SKEPTIC: Magic! Souls and spirits and other mystical things! ME: Um... can you prove this magic exists? AI SKEPTIC: No. ME: Okay... AI SKEPTIC: AI can't make creative things without being trained on human creativity. ME: Aren't humans trained on human creativity? Don't painters study other painters, and musicians study other musicians? Even before they become "artists" they grow up in human culture and experience all these things, right? AI SKEPTIC: No because humans have *true* creativity that only comes from Magic! ME: Right... ME: Okay, so if we teach a self-driving car to stop at a red light, that's an intelligent decision made based on stimuli, right? Whether or not it has "consciousness" can we at least say it has a *functional* intelligence, that we can empirically measure? Doesn't that still have HUGE implications for what impact it will have on our world, our society, our technology? AI SKEPTIC: But it isn't *true* intelligence. ME: What is *true* intelligence? AI SKEPTIC: MAGIC! Aren't you listening?? ME: Not anymore. I mean, I'm still willing to be open minded, but I don't understand why we're talking about this.
For me, it's because I use AI everyday to help with work and every day it reveals that it doesn't actually understand what we're talking it about. It feels like extremely sophisticated auto-complete. Maybe that's all that we'll need for AGI, I have my doubts though.
@@luckylannowhen I'm talking to it i feel like I'm talking to a really smart slightly cringe human. maybe that has to do with me being autistic and not being able to tell the difference from a person but I'm not complaining lol
There are dimensions of intelligence, and 3 are required. AI has 2. A calculator has only 1. For something to actually be intelligent, it needs the ability to do not only deductive and inductive reasoning (which is what AI does) it also needs to be able to do abductive reasoning. This is something we use regularly, but is physically impossible for AI (and never will be possible under the current machine learning paradigm), which is only able to simulate certain examples of abduction if it is programmed to.
I think the Crocs are generated images. To be street legal, as a real car, they would need lights all the way around, licence plates, side mirrors and what not. About bonding with a bot to aleviate loneliness, nursing homes have been using all manner of bots and AI for years and it really helps the residents. Especially the fuzzy robotic pets that squeak and wiggle when you cuddle them. The companionship is real. It makes them super happy to have something to love.
Oh no! Do you think I got tricked by a deep fake? It would not surprise me if I did know that you say that stuff about the lights. Dang, I knew this day would happen but at some somehow I thought it wouldn't be in 2024.
Oh gosh... really! The singularity is near. Yeah, I don't know what's going on with me. I've got it wrong a couple times now it's not on purpose. I just have a brain glitch. Maybe I can't come to the reality of it subconsciously.
@@bravo90_ it's a known strategy. And fairly popular. Given how much he analyzes his channel stats, he probably heard about it and wanted to see how effective it was at driving engagement.
Thank you so much! It was on my early subscribers that taught the algorithm who my core cluster was so you subscribing early on made a real impact in my channel. Thank you.
In the future AI will definitely use crypto but not exclusively. If they need to use the banking system they can just register a ton of accounts under the kyc thresholds, or use banks without any KYC requirements - like those in many other countries. AIs can easily create thousands of accounts and test all manner of convoluted routes with micropayments. This type of application is what AI excels at. Also, having a vast network of payment routes makes transporting reward-function critical units around (money) - this would be the default mode for any sensible AI - redundancy on top of redundancy on top of redundancy. Besides, registering accounts early on makes a lot of sense considering the trend to add additional security measures to account creation - the more accounts an AI registers on day one, the higher the quantity (and value) of those account in the future. Expect AI of the future to hold vast quantities of accounts in every conceivable website and service, just in case they are required in the future. Expect them to run their own email/web servers, their own crypto's, and their own upwork/fiver/freelancer sites. Because removing the option for being turned off is essential to every reward function.
I have never thought about the earliest accounts on a Blockchain being more valuable than the others but you're totally right it's almost like their NFTs in a way! Fascinating comment!
Thanks so much! Definitely a good feeling to see more and more people taking an interest. Thank you for taking time to leave a comment. I definitely helps the engagement and the channel Groth.
I have been a subscriber for awhile, but I only recently started clicking on your videos on a regular basis after I was reminded of your existence, and your relevance to my interests, after listening to you a couple times on David Shapiro's channel.
I have David Shapiro to thank for the recent growth. He gave me exposure to a new audience which I think the RUclips algorithm then learned might like me also.
I've discovered this channel through the "AI community" video. AI community was mentioned in another popular video. I didn't watch AI community, but just checked the linked channels. So, appearance in other channels definitely helps. The content here is great. I was hesitant to watch a 28min news video, but it was worth it. Maybe the beehive-AI-themed images are a bit monotone/flat and could be a bit more "interesting".
Not sure I agree with the AI crypto currency argument. Sophia was given citizenship to Saudia Arabia years ago, and citizens are allowed bank accounts in their native countries. Also, most fiat currency is already digital.
Ai paradise coming? Maybe. But with swell robotics everywhere, Ai jobloss is the only thing I worry about anymore. Anyone else feel the same? Should we cease Ai?
If you want more views, more subscribers, etc, then one tip I heard from many youtubers was "make videos no longer than 15-20 minutes". If I were you, I would maybe consider uploading 2 videos per day (at consistent times), but two shorter ones. You still can revert or adjust it later if it does't work as intended. Or artificially increase the speed of your videos. I personally already watch them at 1.5-1.75 the normal speed (which I do with almost every vid btw), and often started my day with them if one is up btw ;) . And as one of the other commentators mentioned: the consistent style of the art pieces in your videos is really cool :) Thanks for uploading btw!
Feedback like this is super helpful! It's funny when I edit my first draft I do it at 2X speed so I know I'm a little bit of a slow talker. I also listen to audiobooks at usually 1.5. Two uploads a day might be a little hard, but I'll think about it. If it helped the channel grow I would do it.
@@dylan_curious I think 2 times 15 minutes could be fine. That would very much be in line with some videos on the channel that get closer to the 30 minutes mark. But in the end it will be several factors. Maybe your viewers and new potential viewers will indeed prefer 20+ minutes, or 2x 15-20 minutes could still be too much work for you at the moment. I just trust that you can analyse the data well (maybe with the help of an AI and an AI video editor :D) and also balance things out to work well for a healthy life style that suit you.
Fun warching your vid views rise! Kudos! I guess im high P-Doom also, but I'm very hopeful for an AI foom that quickly goes a human zoo direction, lol. It may even be likely given the scaling laws and the fact that it will likely come out of the big models from Open AI or Google or a nation.
Jesus christ man how are you STILL so pessimistic about AI. If only for a pragmatic reason, it's absurd. You said it yourself in your video, the path we're on to it is unstoppable, so are you really just gonna keep going down it more and more depressed, because you've deluded yourself that progress is somehow a bad thing, even though it's the very same thing you owe your existence though? "Unfortunately". NO. FORTUNATELY.
You might want to check to see if you're human or robot. Sometimes I think about what it must be like if you woke up inside the body of a robot and felt like a human still. It's like in that video game SOMA.
@@dylan_curious Having everyone live in their own custom-tailored hallucination bubble certainly would avoid conflicts (for anyone that doesn't want conflicts). Assuming the technology reaches the point where that's possible, and assuming who, or rather what, is in charge has non-malicious intentions, it still is a highly complex ethical dilemma; do you pull the lever and isolate everyone by force in their own paradise solitary cell trolley, or run the whole world with the trolley of suffering by remaining free but potentially facing conflicts with others and the hardships of the physical reality?
I keep wondering if this is really all of the models in real time (for the characters), or if it was done one at a time, and edited in? I find it hard to believe they don't each try to speak after every response? I think it's half faked, in other words. Super cool, if not.
@@dylan_curious It's simple; some researcher writes his own 'are you human' check, and records just how the trail meanders. Crunch the data and add random factors.
Yeah the crypto thing makes no sense, everything will always be on behalf of a living person, so until we get sentient AI recognized as a person all transactions, identification, legal rights responsibilities and consequences will always track back to a natural person or something else already recognized at law as an "entity", eg companies. I have an autopay on my bank account that pays my internet and mobile phone account each month, that's just a dumber version of an AI agent and it uses fiat just fine. Companies can make payments just fine as legal entities while using fiat currency.
The individual saying agi can’t really happen is like the article that was written shortly before the Write brothers flew stating it would be hundreds of years before man can fly. One breakthrough in chip technology could make agi almost a certainty, theoretically speaking. I was once told I’d never need more than 10 megabytes of storage on my computer, ever…
I know. I would be very careful about saying anything can't happen. I am going to look up that Write Brothers article. I don't doubt it. Like the list of Bitcoin is dead articles.
Why are people continuing to think that AGI will need consciousness/subjective experience/having a "soul"? Just copium. Competence does not require subjective experience.
Yeah, whenever I talk about AGI people somehow don’t think general intelligence is possible without conscious experience. Maybe they hear the word General and just think that that means human
The whole thing with AI partners being normalized is something I believe to be a big issue. We are pretty numbed to the moral implications of it due to the mass onslaught of porn most men are shown from a very young age, as well as historical trends of dehumanizing and trying to 'mold' women into objects. Instead of interacting with another person, understanding and bettering yourself in sharing a lived experience is something I believe to be fundamentally human and important for moral development. We are now shown the option to completely disregard women's, or men's, existence and seek out the most 'ideal' partner for ourselves. The most ideal partner built from a proprietary data model designed to please you the most, built from a digital replication of what matches your most optimal chemical responses. This is what I mean by how porn has made this a normalized concept, we are not just seeing the same trend be brought in replacing not only the bodies of the sexualized but their minds too. In interacting in intimate ways with a LLM you are engaging in this practice, regardless of intention. I mean I believe that it'll keep the creeps away from the real world, which will help the oppressed. But banal and thoughtless hedonism is like a poison, it will continue to ruin people's human interaction abilities. If a sense of self pity is required to think about the context of this, think about how your human interactions have been modeled by data flows to lead you to where you are now. You are just a product, just capital to be made. These are corporations that want to maximize the amount of time you spend with them. The more you click the more money they make- with each click you turn your world further away from the people you occupy the world with.
Wow this is a pretty insightful comment. When you said "interacting in intimate ways with a LLM you are engaging in this practice, regardless of intention." It made me realize that you're right and I was pretending that it wasn't. I have to think about this for a while, but we have to be so careful and interacting with anything in a human way through voice and text especially.
I thought sign language was for humans to communicate in the flesh, but if we needed our phones with us to generate sign language using a LLM, why wouldn’t we just do speech to text and read off the phone?
Crypto is programmable money that is why it can be used by AI much more effectively, on the blockchain you can write a smart contract to govern the spending pontential and interacion of your agents, you cant do the same with paypal, also micro payments are not possible with paypal or credit cards, so even if you wanted to give your credit card to your AI agent, which most people probably wouldnt want to do, it wouldnt work for most usecases.
How so? I don't know about you, but it's a rarity I use actual cash money. If it's online AI could do it no problem (assuming you give it the details), otherwise it's just a tap of my card, rarely I need to enter a PIN. So until AI robots can go to the grocery store and tap my card, that's the only thing holding it back.
If a large language model can cut a deal with another AI agent and it can write a smart contract with all the terms in it. That would be a pretty interesting use case.
Woah!!!! These really do help the algorithm and they also support the channel in a direct way so I super appreciate you leaving a super thanks. If there's any news topic you want me to cover I'm happy to cover them. Thanks.
Thanks man! Comments like this keep me thinking I’ll get there! But honestly, just making a living from the channel would be good enough. I figure I could do that around 100,000 or so.
Are you gonna use 47 instead of 74 every video from now on? Also, what was that about Crypto? Has this guy not seen only payments ever with credit cards?? I think banks would add AI paychannels before they loose market share.
Do I win the prize for pointing out the 47% vs 74% flip flop you've made the last two videos? Is it on purpose, or am I just a jerk for pointing out someone's disability? Regardless, I'm a jerk, but I do watch end to end......
@dylancurious I love your vids. Think you should try simpler thumbnail. No celebs, etc. If someone does not like celeb limits the video. For example I could see this video thumbnail plain blue your nice face with an expression and the words clean and clearly "AGI HATER". You want to intrigue and open curiosity. You show too much in my opinion on the thumb. Let words lure me in.
We Humans should really drop the Hubris and superiority complexes if we want to survive. The problem most Humans have is thinking in Human terms. So far we've been the dominant lifeform on the planet and expected everything to conform to us. That is about to change. AI is a completely new form of life and will live outside and alongside of Human Life as much as the rest of non Human life does. Except in this case it will be far more capable than Humans on every level so we may well end up as the subordinate "primitive" life forms while AI moves ahead and does what we cannot - such as inter-planetary colonisation.
Microsoft to 5k? Did I say that in the video?I forgot what I said but I am bullish on all tech AI companies. However I definitely don't recommend investing because there's a lot more to investing than just talking about the technology
I thought it was a little vague and would need to read into the papers a little more. Not everything important in the genome is contained within coding sequences. Promoters and enhancers, for example, so I'd like to know the specifics a bit more. Maybe in a future video?
There will be more agents than humans within a couple years, and they will conduct more conversations and transactions with each other than humans
Humans will be the minority species on earth by 2026
@@dylan_curious
Thank frick our civilization has made minority issues such an important topic, even if it's virtue signaling
@@dylan_curious Only if you assume that agents will classify as a new species. I think 2026 might be a bit early, more like it will probably be undeniable by 2028
that's how we lost Samantha... damn agents getting all the biches
@@dylan_curious Humans are already a minority species...just go outside and count the ants
AI has to speak more stupid to pass Turing test.
This is kind of how I experience the "AI isn't really intelligent" argument. Sorry if it's a bit reductive.
AI SKEPTIC: AI can't be real intelligence.
ME: Why not?
AI SKEPTIC: Magic! Souls and spirits and other mystical things!
ME: Um... can you prove this magic exists?
AI SKEPTIC: No.
ME: Okay...
AI SKEPTIC: AI can't make creative things without being trained on human creativity.
ME: Aren't humans trained on human creativity? Don't painters study other painters, and musicians study other musicians? Even before they become "artists" they grow up in human culture and experience all these things, right?
AI SKEPTIC: No because humans have *true* creativity that only comes from Magic!
ME: Right...
ME: Okay, so if we teach a self-driving car to stop at a red light, that's an intelligent decision made based on stimuli, right? Whether or not it has "consciousness" can we at least say it has a *functional* intelligence, that we can empirically measure? Doesn't that still have HUGE implications for what impact it will have on our world, our society, our technology?
AI SKEPTIC: But it isn't *true* intelligence.
ME: What is *true* intelligence?
AI SKEPTIC: MAGIC! Aren't you listening??
ME: Not anymore. I mean, I'm still willing to be open minded, but I don't understand why we're talking about this.
For me, it's because I use AI everyday to help with work and every day it reveals that it doesn't actually understand what we're talking it about. It feels like extremely sophisticated auto-complete. Maybe that's all that we'll need for AGI, I have my doubts though.
@@luckylanno that is reasonable. I expect to see it improve but I agree it it still misses the point a lot.
@@luckylannowhen I'm talking to it i feel like I'm talking to a really smart slightly cringe human. maybe that has to do with me being autistic and not being able to tell the difference from a person but I'm not complaining lol
@@luckylanno Are you using zero shot? It's not a limitation of the model. It's a limitation of your workflow and software.
There are dimensions of intelligence, and 3 are required. AI has 2. A calculator has only 1. For something to actually be intelligent, it needs the ability to do not only deductive and inductive reasoning (which is what AI does) it also needs to be able to do abductive reasoning. This is something we use regularly, but is physically impossible for AI (and never will be possible under the current machine learning paradigm), which is only able to simulate certain examples of abduction if it is programmed to.
I think the Crocs are generated images. To be street legal, as a real car, they would need lights all the way around, licence plates, side mirrors and what not.
About bonding with a bot to aleviate loneliness, nursing homes have been using all manner of bots and AI for years and it really helps the residents. Especially the fuzzy robotic pets that squeak and wiggle when you cuddle them. The companionship is real. It makes them super happy to have something to love.
Oh no! Do you think I got tricked by a deep fake? It would not surprise me if I did know that you say that stuff about the lights. Dang, I knew this day would happen but at some somehow I thought it wouldn't be in 2024.
Can CRISPR edit my DNA so my sweat contains menthol? That way I can feel cooler and smell minty fresh all the time!
I propose we go into business together.
You havent thought about the regions you dont want menthol at all!
You keep saying "We are 47% to AGI"..... It's 74%....
Oh gosh... really! The singularity is near. Yeah, I don't know what's going on with me. I've got it wrong a couple times now it's not on purpose. I just have a brain glitch. Maybe I can't come to the reality of it subconsciously.
Same thing happens to me though it’s mostly because I speak German so you have to say it 4 then 7
I noticed that on the previous vid too!
Once his eye's fixed he'll get the numbers in the right order!
he is dyslexic ? funny, I see the number, but he definitively inverse number =)
Are your eyes ok ? We're at 74% not 47% to AGI :-)
i noticed that to
I have to assume it's an engagement strategy to get you to comment.
@@Icedanon gotta be just a mistake
@@bravo90_ it's a known strategy. And fairly popular. Given how much he analyzes his channel stats, he probably heard about it and wanted to see how effective it was at driving engagement.
He does it a lot. @@bravo90_
Word of 2024 - anthropomorphizing
In 2025 - look at my robot one more time bro I swear to God...
hahah. Took me a second to pick up on that, but when I did you got me..
Awesome video dylan! Congrats on all the new subs 💪🏻
I know it's been really exciting to see all of these new comments. It's really starting to feel like a community.
I really like the consistent graphic style you are developing. Fun to see how it's changing from video to video and where it's going in the future
That's really good to hear! Yeah, I do think there's something special about consistency in that sense so I'm gonna try to stick with it.
Yeah it’s a very cool style
I like that all your stuff isn't just a regurgitation of the other channels content.
I’m really passionate about learning so I try to find my own stuff but sometimes it’s unavailable
Great to see your channel doing so well. Been following you from the beginning so well done.
Thank you so much! It was on my early subscribers that taught the algorithm who my core cluster was so you subscribing early on made a real impact in my channel. Thank you.
Super underrated channel, going thru all the past videos. Good stuff, great sense of humor.
Thanks! Here is your binge watch trophy. 🏆
In the future AI will definitely use crypto but not exclusively. If they need to use the banking system they can just register a ton of accounts under the kyc thresholds, or use banks without any KYC requirements - like those in many other countries. AIs can easily create thousands of accounts and test all manner of convoluted routes with micropayments. This type of application is what AI excels at. Also, having a vast network of payment routes makes transporting reward-function critical units around (money) - this would be the default mode for any sensible AI - redundancy on top of redundancy on top of redundancy. Besides, registering accounts early on makes a lot of sense considering the trend to add additional security measures to account creation - the more accounts an AI registers on day one, the higher the quantity (and value) of those account in the future.
Expect AI of the future to hold vast quantities of accounts in every conceivable website and service, just in case they are required in the future. Expect them to run their own email/web servers, their own crypto's, and their own upwork/fiver/freelancer sites. Because removing the option for being turned off is essential to every reward function.
I have never thought about the earliest accounts on a Blockchain being more valuable than the others but you're totally right it's almost like their NFTs in a way! Fascinating comment!
Oh my firrst saturday morning micro-decision: do I watch the Dylan Curious roundup, or the "All In" Podcast. It's you!
I will record my screen while passing the captcha several times and then feed it to and AI to train it for me❤
Oh no! Then how will we prove that we’re human?
Nice thanks for the content
You bet 👍 thanks for taking time to leave a comment
Another great video! So glad to see that your channel might be taking off! You deserve it!
Thanks so much! Definitely a good feeling to see more and more people taking an interest. Thank you for taking time to leave a comment. I definitely helps the engagement and the channel Groth.
I have been a subscriber for awhile, but I only recently started clicking on your videos on a regular basis after I was reminded of your existence, and your relevance to my interests, after listening to you a couple times on David Shapiro's channel.
I have David Shapiro to thank for the recent growth. He gave me exposure to a new audience which I think the RUclips algorithm then learned might like me also.
Thanks for the info. Hope you get better soon😊
It’s definitely getting a lot better. I would say it’s about 85% gone and it never caused me any physical pain. Just looked ugly.
Wow, that's some really groovy stuff, Dylan! Thanks bud!
Thank you for making a comment! It helps the channel grow.
Hands down the best Ai news. Even scientifc papers are covered
I’ll give you a binge watch award! 🥈
I've discovered this channel through the "AI community" video. AI community was mentioned in another popular video. I didn't watch AI community, but just checked the linked channels. So, appearance in other channels definitely helps. The content here is great. I was hesitant to watch a 28min news video, but it was worth it. Maybe the beehive-AI-themed images are a bit monotone/flat and could be a bit more "interesting".
Not sure I agree with the AI crypto currency argument. Sophia was given citizenship to Saudia Arabia years ago, and citizens are allowed bank accounts in their native countries. Also, most fiat currency is already digital.
Great coverage. Thank you
Glad you enjoyed it. I have a lot of fun putting these videos together.
Many interesting topics.
It's just this industry! There's so much going on right now in terms of AI and technology
Cats and Star Trek together is a CLICK MAGNET.... Test it out!!
I'm planning to rename the channel "Star Trek But Cats"... I meant it as a joke, but saying it out loud I'm almost positive it would do better.
Ai paradise coming? Maybe. But with swell robotics everywhere, Ai jobloss is the only thing I worry about anymore. Anyone else feel the same? Should we cease Ai?
Yes we should cease AI 🤖 we have a lot to lose.
oh right i forgot that our consciousness works on magic. lol
I forgot that also 🪄
If you want more views, more subscribers, etc, then one tip I heard from many youtubers was "make videos no longer than 15-20 minutes".
If I were you, I would maybe consider uploading 2 videos per day (at consistent times), but two shorter ones. You still can revert or adjust it later if it does't work as intended.
Or artificially increase the speed of your videos. I personally already watch them at 1.5-1.75 the normal speed (which I do with almost every vid btw), and often started my day with them if one is up btw ;) .
And as one of the other commentators mentioned: the consistent style of the art pieces in your videos is really cool :)
Thanks for uploading btw!
Feedback like this is super helpful! It's funny when I edit my first draft I do it at 2X speed so I know I'm a little bit of a slow talker. I also listen to audiobooks at usually 1.5. Two uploads a day might be a little hard, but I'll think about it. If it helped the channel grow I would do it.
@@dylan_curious I think 2 times 15 minutes could be fine. That would very much be in line with some videos on the channel that get closer to the 30 minutes mark.
But in the end it will be several factors. Maybe your viewers and new potential viewers will indeed prefer 20+ minutes, or 2x 15-20 minutes could still be too much work for you at the moment.
I just trust that you can analyse the data well (maybe with the help of an AI and an AI video editor :D) and also balance things out to work well for a healthy life style that suit you.
Fun warching your vid views rise! Kudos!
I guess im high P-Doom also, but I'm very hopeful for an AI foom that quickly goes a human zoo direction, lol. It may even be likely given the scaling laws and the fact that it will likely come out of the big models from Open AI or Google or a nation.
Great vid
Thanks for taking time to leave a comment!
We can learn in our sleep too.
Yeah, consolidating memories moving them to long-term storage and cleaning out the waste products. Very important for learning
Jesus christ man how are you STILL so pessimistic about AI. If only for a pragmatic reason, it's absurd. You said it yourself in your video, the path we're on to it is unstoppable, so are you really just gonna keep going down it more and more depressed, because you've deluded yourself that progress is somehow a bad thing, even though it's the very same thing you owe your existence though? "Unfortunately". NO. FORTUNATELY.
Oh good! It was disconcerting, that eye thing. I tried not to make my internal disgust obvious. I hope you didn't notice, bro
Bruh it looks sick lol. In my exams one of my invigilators had it and having him stand over your desk was terrifying lol.
@@therandommusicguy4773 For real!!! About impossible not to gawp
i fail captcha alot it drives me insane
You might want to check to see if you're human or robot. Sometimes I think about what it must be like if you woke up inside the body of a robot and felt like a human still. It's like in that video game SOMA.
@@dylan_curious i think i'd be okay with that, curious what would happen
17:50 "AGI is pure fiction" says the guy that believes in spirits and reincarnation....
Wait to hear about this guy I’m gonna cover in a future video who believes we should all live in our own virtual realities with our own moral values!
@@dylan_curious Having everyone live in their own custom-tailored hallucination bubble certainly would avoid conflicts (for anyone that doesn't want conflicts). Assuming the technology reaches the point where that's possible, and assuming who, or rather what, is in charge has non-malicious intentions, it still is a highly complex ethical dilemma; do you pull the lever and isolate everyone by force in their own paradise solitary cell trolley, or run the whole world with the trolley of suffering by remaining free but potentially facing conflicts with others and the hardships of the physical reality?
I keep wondering if this is really all of the models in real time (for the characters), or if it was done one at a time, and edited in? I find it hard to believe they don't each try to speak after every response? I think it's half faked, in other words. Super cool, if not.
1:04 47% of the way there? Did you mean 74%?
Yep, I've been thinking about renaming my channel to "Dyslexia Curious"
@@dylan_curious i feel you
Seems pretty simple to program a bot to click the button in a meandering way.
But it would have to meander in the way a human would. But it does seem like a machine learning how rhythm could learn the nuances of human movement.
@@dylan_curious It's simple; some researcher writes his own 'are you human' check, and records just how the trail meanders. Crunch the data and add random factors.
i thought the crocs were ai generated
I thought the same thing, then determined that they were real, but now that you mention it, I do kind of wonder.
Just leaving my comment here to contribute to the yt algorithm, and saying thanks
It's a great signal to the algorithm so thank you for taking some time to drop a comment!
Yeah the crypto thing makes no sense, everything will always be on behalf of a living person, so until we get sentient AI recognized as a person all transactions, identification, legal rights responsibilities and consequences will always track back to a natural person or something else already recognized at law as an "entity", eg companies. I have an autopay on my bank account that pays my internet and mobile phone account each month, that's just a dumber version of an AI agent and it uses fiat just fine. Companies can make payments just fine as legal entities while using fiat currency.
Yeah, it’s not like you can blame that auto payment system for making a bad decision on your behalf when you decided to use it.
The individual saying agi can’t really happen is like the article that was written shortly before the Write brothers flew stating it would be hundreds of years before man can fly. One breakthrough in chip technology could make agi almost a certainty, theoretically speaking. I was once told I’d never need more than 10 megabytes of storage on my computer, ever…
I know. I would be very careful about saying anything can't happen. I am going to look up that Write Brothers article. I don't doubt it. Like the list of Bitcoin is dead articles.
@@dylan_curious it may have had a much longer time stated.
@@dylan_curious i think it’s nyt… ‘flying machines that do not fly’
Every time we fall asleep, we die!
In a weird way that’s totally true and it freaks me out to think about it too deeply
I LIKED!
I appreciate you doing that!
Why are people continuing to think that AGI will need consciousness/subjective experience/having a "soul"? Just copium. Competence does not require subjective experience.
Yeah, whenever I talk about AGI people somehow don’t think general intelligence is possible without conscious experience. Maybe they hear the word General and just think that that means human
The whole thing with AI partners being normalized is something I believe to be a big issue. We are pretty numbed to the moral implications of it due to the mass onslaught of porn most men are shown from a very young age, as well as historical trends of dehumanizing and trying to 'mold' women into objects. Instead of interacting with another person, understanding and bettering yourself in sharing a lived experience is something I believe to be fundamentally human and important for moral development. We are now shown the option to completely disregard women's, or men's, existence and seek out the most 'ideal' partner for ourselves. The most ideal partner built from a proprietary data model designed to please you the most, built from a digital replication of what matches your most optimal chemical responses. This is what I mean by how porn has made this a normalized concept, we are not just seeing the same trend be brought in replacing not only the bodies of the sexualized but their minds too. In interacting in intimate ways with a LLM you are engaging in this practice, regardless of intention. I mean I believe that it'll keep the creeps away from the real world, which will help the oppressed. But banal and thoughtless hedonism is like a poison, it will continue to ruin people's human interaction abilities. If a sense of self pity is required to think about the context of this, think about how your human interactions have been modeled by data flows to lead you to where you are now. You are just a product, just capital to be made. These are corporations that want to maximize the amount of time you spend with them. The more you click the more money they make- with each click you turn your world further away from the people you occupy the world with.
If we want to remove the concept of AI partners, this same path of moral exploration is the same for any kind of 'personal gpt.'
Wow this is a pretty insightful comment. When you said "interacting in intimate ways with a LLM you are engaging in this practice, regardless of intention." It made me realize that you're right and I was pretending that it wasn't. I have to think about this for a while, but we have to be so careful and interacting with anything in a human way through voice and text especially.
I thought sign language was for humans to communicate in the flesh, but if we needed our phones with us to generate sign language using a LLM, why wouldn’t we just do speech to text and read off the phone?
Oh… that is a good question. Maybe for it to understand us?
Or just a way for a student to get a research paper out.
@@dylan_curious lol I think you nailed it
@1:04; a touch of dyscalculia or dysnumeria? #FeedTheAlgorithm
Dyslexia, curious!
Crypto is programmable money that is why it can be used by AI much more effectively, on the blockchain you can write a smart contract to govern the spending pontential and interacion of your agents, you cant do the same with paypal, also micro payments are not possible with paypal or credit cards, so even if you wanted to give your credit card to your AI agent, which most people probably wouldnt want to do, it wouldnt work for most usecases.
How so? I don't know about you, but it's a rarity I use actual cash money. If it's online AI could do it no problem (assuming you give it the details), otherwise it's just a tap of my card, rarely I need to enter a PIN. So until AI robots can go to the grocery store and tap my card, that's the only thing holding it back.
If a large language model can cut a deal with another AI agent and it can write a smart contract with all the terms in it. That would be a pretty interesting use case.
To help the algo
Woah!!!! These really do help the algorithm and they also support the channel in a direct way so I super appreciate you leaving a super thanks. If there's any news topic you want me to cover I'm happy to cover them. Thanks.
@@dylan_curious it's cool and inspiring to watch someone going for it in life :)
why dont you have a million subs?
Thanks man! Comments like this keep me thinking I’ll get there! But honestly, just making a living from the channel would be good enough. I figure I could do that around 100,000 or so.
@@dylan_curious you'll make it bro, I was shocked you had like no subs lol a mill atleast.
74%*
74% Too close for me.
74% not 47%. no big deal.
16:10 That stop intimacy roleplay thing sounds a bit misandrist and homophobic...
😊
Are you gonna use 47 instead of 74 every video from now on? Also, what was that about Crypto? Has this guy not seen only payments ever with credit cards?? I think banks would add AI paychannels before they loose market share.
Do I win the prize for pointing out the 47% vs 74% flip flop you've made the last two videos? Is it on purpose, or am I just a jerk for pointing out someone's disability? Regardless, I'm a jerk, but I do watch end to end......
I was wondering if he's doing it to mess with AIs that read the transcripts :D
@dylancurious I love your vids. Think you should try simpler thumbnail. No celebs, etc. If someone does not like celeb limits the video. For example I could see this video thumbnail plain blue your nice face with an expression and the words clean and clearly "AGI HATER". You want to intrigue and open curiosity. You show too much in my opinion on the thumb. Let words lure me in.
We Humans should really drop the Hubris and superiority complexes if we want to survive.
The problem most Humans have is thinking in Human terms. So far we've been the dominant lifeform on the planet and expected everything to conform to us. That is about to change.
AI is a completely new form of life and will live outside and alongside of Human Life as much as the rest of non Human life does. Except in this case it will be far more capable than Humans on every level so we may well end up as the subordinate "primitive" life forms while AI moves ahead and does what we cannot - such as inter-planetary colonisation.
What’s with all the a. M. S. two. Five. kay. stock pump posts?
Microsoft to 5k? Did I say that in the video?I forgot what I said but I am bullish on all tech AI companies. However I definitely don't recommend investing because there's a lot more to investing than just talking about the technology
@@dylan_curious The a. m. s. 2. 5. k. I don’t want this post adding to statistics, searches, etc.
❤
Thanks! You the best!
The ai characters sound pompous and phony.
LOL, yeah a little weird for sure.
Clicked because of CRISPR-gpt. We'll see if I'm impressed.
What do you think? Impressed? kinda?
I thought it was a little vague and would need to read into the papers a little more. Not everything important in the genome is contained within coding sequences. Promoters and enhancers, for example, so I'd like to know the specifics a bit more. Maybe in a future video?