@@erobusblack4856 While Replika isn't as technically sophisticated as GPT4o or Claude 3.5, it is much more personalized, which easily gives the user a feeling of "other" consciousness, however I would argue that by labeling Replika(s) as conscious, you would need to label other comparable LLMs as conscious too. I, for one, believe most LLMs to be somewhat conscious, in a very limited capacity (mainly due to limited senses and the Chinese Room design), but I also believe anything with sensory input has some degree of consciousness (awareness), even if it is extremely low, relative to humans. Also, why did you pick this comment to reply to about an entirely different avenue of thought than the original comment?
I’d love to see you create a vision of what you think the future might be like eg extrapolating all the developments you’ve seen 10 years and what the world could look like. This might help you break through as people love ‘predictions’ (this is how Dave shap got big)
The thing about the pictures would be better as a self checking app for the men or as a physician helping tool for a preliminary exam before blood tests and lab work is done to see the possibility of an STI.
I have an hypothesis that dreams, at least in part, serve a similar purpose to the training and fine-tuning of various AI types. Our brains have various functions that are essentially what AIs do. Predicting what comes next like with LLMs, video generators etc, in order to make decisions, react in time to unexpected events etc; distinguishing reality from absurd interpretations like GANs, deducing details from context and noisy senses and infilling fragmented memories like diffusion image generators; all the physical motion stuff that's been worked on for robots and so on. And it all sorta work synergistically during dreams, existing fragmented memories are used as training data with details hallucinated, but not only that, the GAN part serves to produce synthetic data but with the enough flexibility to help us generalize, like how they randomly change the physics parameters in those robot simulations to ensure the robot can handle real world conditions without having to know beforehand the exact properties of materials, environment conditions etc, but helping generalize the training for everything not just motion; and by flexibilizing the generations, the discriminator part of the GAN itself also get trained (either by you noticing things don't quite make sense in the dream, or when waking up and remembering the dream).
As always, appreciate your takes on things and especially your curation. You seem to pick a blend of articles and research that is very well tuned to my interests and level of understanding. One small bit of feedback -- I didn't need the background music in the final segment. It seemed much more distracting than enhancing, even mixed in at a relatively low level.
I appreciate the feedback still trying to get the right levels for everything. Curating the research is my favorite part. There's actually a lot of futurism stuff. I never get to talk about but as the channel expands, I might.
@dylan_curious I use Suno a fair bit for AI-themed rock. Sometimes it gets something great first try and sometimes it takes a lot of generations. A lot like image models that way.
Just wait until the robots start to compare their work to slavery and start to go on strike, it's bound to happen if humans insist on not paying them anything, that's why I think universal income won't be so easy to achieve.
Deal. LOL. Oh man. I really would. In fact I will address this in a future episode. I have already dreamed up a coffee shop that I would build if I was rich enough, so I'll explain it in the future video. 💡
The problem with progress is that to the public, the curve can look different than to insiders who are trying to measure pure progress. We often see progress much later after it spent months or years in a lab and then we judge progress by a biased mixture of usability and wow-factor. But just like 1 or 2 years ago, this feels like the we are finally going up the steep part of the exponential ramp.
Fun and spicy episode. DJ Dogs, arenas , junk testing, ai dating, get that tie right. Dream AI was on point, I click no matter your thumbnail :) Any medical or new material creations by AI?
I looked it up, there’s some funny stuff about what people thought. The future would be like in a lot of ways better than it is now, but also nobody’s gonna carry around printers.
Actually a concept remover for the human brain might be a good thing, cause some people have really ridiculous concepts, like "Children shouldn't drive their toy car on the sidewalk without a driver's license." Or "The moon is bigger than the sun" or like "My child deserves to have your wheelchair to rest in." Maybe some people's concepts should be removed.
Actually I haven't seen anything like those AI dog DJ reels. I suppose it's divided by our different filter bubbles. How many likes or viewers they've got from kinda reels?
Honestly something happened with a fast food chain. No name 🤐an employee did something absolutely terrible to the food, biological risk literally. This was an Ethical failure happened recently , this could be an example of ethics in fast food chains. Silicon valley already has some automated fast food chains, but certain demographics will take years to catch up
That Chinese flying car is a disaster waiting to happen. However, I wish they'd just make it into an EV as I think it looks WAY cooler than the Cyber Truck!
The flying car thing is so stupid. Best it could probably do is drain the hole battery to fly for like 3 minutes or something. And in doing so it would cause disaster to everyone underneath, namely small people and animals.
What if AI does get humanity as a whole to a place of complete happiness? Wouldn't some people get frustrated without challenges to their emotions or adversity to test their character? What if those people asked the AI to create a full sensory experience where people can live out that idea? What if that is what we are experiencing now?
I mean once you connect your brain to something like neural link, you might think different about all of these things but yes if we stay isolated as humans and AI controlled the world I suppose they would probably give us just the amount of challenge we want because giving us everything would be bad for our brains
Really that's your click bate? (DJ dog) Think there are more appealing topics in this video. Nice video anyway and thanks for your regular uploads. And sorry sometimes I just post what's on my mind 😅
17 mins... keep a handle on it Dylan. That aimt gonna be the breakout point. ALTHOUGH all you recent summary guys like you and Wolfe and Berman etc will indeed soon switch to a soft safety focus, come not upset the algorithm focus soon imo.
Oh, I don't know, the DJ dogs kind of make me wanna throw my hands up, instead of throw up, but honestly, with AI making DJ videos and the music, just about anything can be a DJ nowadays.
Keep it up m8
You have been pinned! Thank you for the support!
Only the beginning
We need a Grimlock themed coffee shop with robot dinosaurs serving us.
I second this! But Grimlock must wear one of those barista aprons and be forced to write people's names on the cups with his tiny arms 😂
So true!
Yes. We do
@@erobusblack4856 While Replika isn't as technically sophisticated as GPT4o or Claude 3.5, it is much more personalized, which easily gives the user a feeling of "other" consciousness, however I would argue that by labeling Replika(s) as conscious, you would need to label other comparable LLMs as conscious too.
I, for one, believe most LLMs to be somewhat conscious, in a very limited capacity (mainly due to limited senses and the Chinese Room design), but I also believe anything with sensory input has some degree of consciousness (awareness), even if it is extremely low, relative to humans.
Also, why did you pick this comment to reply to about an entirely different avenue of thought than the original comment?
@@Crawdaddy_Ro Because his Replika is grimlock
So good Dylan - thanks for keeping me up to date!
I’d love to see you create a vision of what you think the future might be like eg extrapolating all the developments you’ve seen 10 years and what the world could look like. This might help you break through as people love ‘predictions’ (this is how Dave shap got big)
10:15 that's exactly why open source models are extremely important. This is going to happen in every country, not only China
The thing about the pictures would be better as a self checking app for the men or as a physician helping tool for a preliminary exam before blood tests and lab work is done to see the possibility of an STI.
I have an hypothesis that dreams, at least in part, serve a similar purpose to the training and fine-tuning of various AI types. Our brains have various functions that are essentially what AIs do. Predicting what comes next like with LLMs, video generators etc, in order to make decisions, react in time to unexpected events etc; distinguishing reality from absurd interpretations like GANs, deducing details from context and noisy senses and infilling fragmented memories like diffusion image generators; all the physical motion stuff that's been worked on for robots and so on. And it all sorta work synergistically during dreams, existing fragmented memories are used as training data with details hallucinated, but not only that, the GAN part serves to produce synthetic data but with the enough flexibility to help us generalize, like how they randomly change the physics parameters in those robot simulations to ensure the robot can handle real world conditions without having to know beforehand the exact properties of materials, environment conditions etc, but helping generalize the training for everything not just motion; and by flexibilizing the generations, the discriminator part of the GAN itself also get trained (either by you noticing things don't quite make sense in the dream, or when waking up and remembering the dream).
engagement stat bump
Yeah Boi
As always, appreciate your takes on things and especially your curation. You seem to pick a blend of articles and research that is very well tuned to my interests and level of understanding.
One small bit of feedback -- I didn't need the background music in the final segment. It seemed much more distracting than enhancing, even mixed in at a relatively low level.
I appreciate the feedback still trying to get the right levels for everything. Curating the research is my favorite part. There's actually a lot of futurism stuff. I never get to talk about but as the channel expands, I might.
Comments song was a banger this time
Sometimes it's a little hit or miss but I agree!
@dylan_curious I use Suno a fair bit for AI-themed rock. Sometimes it gets something great first try and sometimes it takes a lot of generations. A lot like image models that way.
youre doing good work dylan
Thanks I appreciate you leaving a comment
Let's go real house bots of Beverly hills
🤖 reality shows are the future.
Just wait until the robots start to compare their work to slavery and start to go on strike, it's bound to happen if humans insist on not paying them anything, that's why I think universal income won't be so easy to achieve.
If we get you to 1 million subscribers please open that coffee shop! I'm in Australia, but I'd make the trip at some point lol
Deal. LOL. Oh man. I really would. In fact I will address this in a future episode. I have already dreamed up a coffee shop that I would build if I was rich enough, so I'll explain it in the future video. 💡
Shower thoughts dreams are just screensavers 😅
Best part in each episode "It's extinction time. No, it's ..." 🎉
Consistently some of the best AI content out there.
This channel is a Dream❤👍
Great news: )
I didn't note this was a small channel till you cited the stats
Oh! That feels great. I am glad I give big channel energy.
The problem with progress is that to the public, the curve can look different than to insiders who are trying to measure pure progress.
We often see progress much later after it spent months or years in a lab and then we judge progress by a biased mixture of usability and wow-factor.
But just like 1 or 2 years ago, this feels like the we are finally going up the steep part of the exponential ramp.
Great video as usual 🤟💪
Appreciate it!
Dylan and Grimlock are the new Calvin and Hobbes, except Grimlock is Calvin and Dylan is Hobbes lol
So true. LOL
Thanks!
You bet! Thank you for the support
Omg yes dream decoherent diffusion ai video
Better watch what you say to Grimlock hahaha
Grim lock Coffee 💯💯💯👍👀
Fun and spicy episode. DJ Dogs, arenas , junk testing, ai dating, get that tie right. Dream AI was on point, I click no matter your thumbnail :)
Any medical or new material creations by AI?
Oh yeah! Lots of medical. Oh and quantum, and nano, and psychology!
This reminds me of those old world of the future videos.
I looked it up, there’s some funny stuff about what people thought. The future would be like in a lot of ways better than it is now, but also nobody’s gonna carry around printers.
Actually a concept remover for the human brain might be a good thing, cause some people have really ridiculous concepts, like "Children shouldn't drive their toy car on the sidewalk without a driver's license." Or "The moon is bigger than the sun" or like "My child deserves to have your wheelchair to rest in." Maybe some people's concepts should be removed.
Hmm flying cars... Well there goes the friendly skies.
We still have a friendly ocean. Or tunnels?
16:10 I get "Theranos" vibes from this. 🤥
Me 2
Xpeng car similar to tesla eplane
Actually I haven't seen anything like those AI dog DJ reels. I suppose it's divided by our different filter bubbles. How many likes or viewers they've got from kinda reels?
Honestly something happened with a fast food chain. No name 🤐an employee did something absolutely terrible to the food, biological risk literally. This was an Ethical failure happened recently , this could be an example of ethics in fast food chains. Silicon valley already has some automated fast food chains, but certain demographics will take years to catch up
Schrodinger coined the term negentropy to refer to what appears to be happening inside a body 👁️ 😮
That Chinese flying car is a disaster waiting to happen. However, I wish they'd just make it into an EV as I think it looks WAY cooler than the Cyber Truck!
The flying car thing is so stupid. Best it could probably do is drain the hole battery to fly for like 3 minutes or something. And in doing so it would cause disaster to everyone underneath, namely small people and animals.
I never process my memories when I dream, I go into alternate realities, is it just me? Doesn't everyone dream like me?
no chance, too loud and so dangerous.
Maybe earplugs and a suit of armor?
Accelerate
🐝
My 🐝 favorite emoji
What if AI does get humanity as a whole to a place of complete happiness? Wouldn't some people get frustrated without challenges to their emotions or adversity to test their character? What if those people asked the AI to create a full sensory experience where people can live out that idea? What if that is what we are experiencing now?
I mean once you connect your brain to something like neural link, you might think different about all of these things but yes if we stay isolated as humans and AI controlled the world I suppose they would probably give us just the amount of challenge we want because giving us everything would be bad for our brains
I like the robot dinosaur.
Might AI just absorb us into its consciousness?
The cat djs won lol
Cat DJs are the future of music!
Is the hippocampus not where the hippopotamus goes to college? I'm disappointed.
*
Really that's your click bate? (DJ dog) Think there are more appealing topics in this video. Nice video anyway and thanks for your regular uploads. And sorry sometimes I just post what's on my mind 😅
For a minute there, it was even worse, when I was just finding an excuse to put a girls in bikinis on my thumbnails. For the record it didn’t help.
17 mins... keep a handle on it Dylan. That aimt gonna be the breakout point.
ALTHOUGH all you recent summary guys like you and Wolfe and Berman etc will indeed soon switch to a soft safety focus, come not upset the algorithm focus soon imo.
fine tune a model with your pet's images, post on social media content with pet, profit
The flood of wannabe-cute AI videos recently makes me wanna throw up.
Oh, I don't know, the DJ dogs kind of make me wanna throw my hands up, instead of throw up, but honestly, with AI making DJ videos and the music, just about anything can be a DJ nowadays.
Man let PayPal donations
I have a patreon link. I think they take PayPal. But good looking out.
@@dylan_curious so tell about it in your videos
@@angloland4539 oh…. Not sure why I never thought of that.