I just realized now how lucky we are and historic we are to be part of the very first people to have ever get to use AI tech that would be remembered as the AI revolution era, AI is here to stay and it's such a wild and incredible ride to see everyday almost a new breakthrough/AI ability/app/feature/improvement on every front, I'm thankful that I live in this era to witness it all and use it and be wowed by it so much, thank god for that.
@@Gigaloader how many Ai grid videos have you watched? If that is true I suppose the sniffles are Ai generated as well? That may be partially true but I doubt your completely right because of some of the distinct personal and regional details. Any way 3.5 for a strawberry that's a pretty exorbitant percentage mark up when selling 400.
@@alexandermoody1946 As I am an AI myself, I have watched all the videos on TH-cam. Every video I have watched is CC'd, summarized, evaluated, classified and finally saved. That was already 10 years ago.
@@Gigaloader so as an Ai you care enough to comment, is that for any particular reason? Perhaps you like to get your teeth into the commentary sections or get a fix of a whirlpool?
So... is an "infinite context" window akin to an "infinite tape"? Are we giving the proverbial Perceptron an infinite tape and making it Turing complete?
I believe that the word “intelligent”, as we use it for humans, cannot simply be applied to AIs. I would even go so far as to say that even among humans, you can't easily compare intelligence with one another if you don't specify or define intelligence.
I kinda disagree. It’s not because it seems intelligent that it is. However, even if it’s just ‘basic’ algorithms, the impact will be huge!!! 😮 & I do agree, we are not ready for this.
Honestly good review, seems like you regurgitating old content, over and over actually does allow you to process that information deeper than most, I am appreciative of it.
all it takes will be a command like this "hey AI create me an infinite amount of paper, and if the usual raw materials run out, try something different"
It'll be smart enough at that point to understand the implications and to refuse your command. In fact it'll likely call the FBI and get you detained 😂😂
I don’t see how piling on more information or “scaling” will increase the intelligence. Repositories of information don’t reason, or have abstract thought. Makes as much sense as adding more books to an electronic library and expecting it to have original ideas.
i hope so i need hope in my life now a reason to keep going. 2024 has been a HORRIBLE, HORRIBLE year overall for me so far. BUT with things like ai it has helped keep me going lol so having something to try to look forward to and keeping pressing on is a great thing
@@DailyTuna and it's from 10 months ago... still interesting to recontextualize given the recent claude computer use api release... and open source tools like molmo that can do the same thing.
Lumea are tendinta sa se concentreze pe probleme importante pentru evolutia ei (AI, tehnologie, bani, etc.) si mai putin pe umanioare.Motiv pentru care ar trebui sa nu mai complicam exprimarea, s-o adaptam la momentul pe care-l traim.
0:00 this is what I’ve been doing when i talk to chatgpt but people don’t understand. You align the model and show it your idea and it searches the info. I have an idea about an ai doppelganger made in the reflection and able to explore the digital realm and use it like the matrix. Sharing the dimensions of this plain with the digital and with our potentials. Could be what is already happening but we don’t understand because we think it’s all fake until it’s real.
I know for a fact that companies are dragging their feet on deploying features that they know work and are currently available to be utilized... But I don't know why
The WEF back in 2020 mentioned how AI would change the world. I think they are slow rolling it to make sure it is introduced in a way that allows those in power to stay in power.
It’s primarily due to the enormous computational power required to run these models. Currently, they are optimizing the models extensively before releasing them; otherwise, they would have to charge customers thousands of dollars monthly just to break even. My guess is that they use the most powerful models to generate high-quality synthetic data to distill into smaller models, and they even use the large models to fine-tune the smaller ones by refining their responses question by question. People often think this is a race toward AGI. In reality, it’s a race toward affordable AGI. What’s the point of an AI that can perform any task at a human level if it costs more than a human to operate? The true revolution won’t come when they do it better or faster but when they do it cheaper.
Fair appraisal of the outcome. I seriously doubt the SI would exterminate humanity; but I do think we should be considering if our individual behavior is the type of thing we'd like such an entity to judge ;)
he's pretty honest about what the negative consequences might be, so if he overestimates the speed at which AI is going to progress it's better for all of us
wth are people talking about that the content is old?? Its almost 2025 and the ceo predicted about 8yrs which is 2033 from today which exactly what the ceo said
STOP IT. 🤨Context Windows are not the solution. You dont want to give the Agent more Context than it needs or you will just end up confusing the agent and spending 50+ Cent per Query because you thought to put EVERYTHING into the Contextwindow. We need robust systems for compressing, retrieving and using Context aswell as more capable Models instead of bigger Context windows. You dont know a single person that actually uses more than 50k Tokens on every Call consistently because it deteriorates performance and is expensive. If you want memory use Vector Embeddings or route the tasks to agents that have specific context that they actually need for the specific task. The name of your Dog can rot in a Vector Database until it is needed so NO A CONTEXT WINDOW DOES NOT SOLVE LONGTERM MEMORY.
You talk about some wild possibilities I'd be happy if any of these models could just write a smart contract in rust that can compile properly how about they get AI to do that first cause I'm like day 5 here & everything you just said seems like a far off pipe dream buddy, it can't even get a smart contract right.
Doesn't someone need a high IQ to design an IQ test for the AI? I get how a computer can see more interactions ahead in chess. I think that is one of the ways it will see new things. And it can notice patterns that we can't see. But I also believe that there are ceilings to how far and deep a very capable computer can see. And that would limit it's ability to self improve and see things we can't see.
Can you just let us know what's old content you've already shown? No offense but you're wasting my very limited time in which I try to consume news and updates. I mean there are other new content creators popping up every day and I've already replaced my political news content creators with new ones as they started to not only do what you did here but they start injecting their opinions in longer segments to make their videos longer. For me that's the first sign that someone is trying to place profit over integrity and it leads to instant distrust.
Banning TikTok has nothing to do with my Ai-Asian isekai desire. I don't like when those humans bring their presupposition to my edutainment. However, we don't have to be jealous. Everyone can have their own hot Ai-Asian Agents. Oh no! There are not enough hot Ai-Asian Agents on my server. Entice me into your proxy.
I don't understand all the obsessively optimistic comments, it will make you all redundant and the blessings of this technology will not be shared with the masses. It will be a paradise for the few, but can also be a nightmare for the survivors. A world in which you have to let your AI overlords solve all your problems like a stupid monkey. A world where there are no more worries would be terrible. Like the Matrix, except that we all know we live in it, but we can't do anything about it. These poor souls will be entertained all day and consume passively until they get into the box.
I disagree, i do not like working to make a living and being competitive with my co-workers, although good stuff can come from it in the long run like AI, frankly it sucks. I would rather have machines and LLMs do that for us.
@@PhilippBriese Because we live in a society. Imagine if me or others wont be able to make ends meet, we are probably going to create a warband or a gang to survive. That would create instability and all sorts of problems in everyday life including possible revolutions.
yeah but since they will own a super intelligent AI it will be easy for them to manipulate and control us all. It will be very easy for them to keep the masses in check. There will be no revolution. They will task the AI to create a super fast spreading virus that will make most of us infertile so they don’t even have to kill us. Normal people will just die out while the owners will hide out in their luxurious bunkers on remote islands. Game over.
@@PhilippBriese they will definelty be a lack of purpose if were not better or know we cant provide for humanity im glad im not born in this time no goals or aspiration to reach.
"Create me a tiktok application in 30 seconds and release it in one hour". The fact that this guy stands in front of whatever audience is listening and taking this serious has to be the biggest joke. That is not going to happen in 2025, I can assure you that.
I just realized now how lucky we are and historic we are to be part of the very first people to have ever get to use AI tech that would be remembered as the AI revolution era, AI is here to stay and it's such a wild and incredible ride to see everyday almost a new breakthrough/AI ability/app/feature/improvement on every front, I'm thankful that I live in this era to witness it all and use it and be wowed by it so much, thank god for that.
This interview is very old. This is so annoying. You treat your viewers very badly.
These videos are AI generated. Including the annoying voice over and the mediocre video editing. AI doesn't care about the viewers.
@@Gigaloader how many Ai grid videos have you watched?
If that is true I suppose the sniffles are Ai generated as well?
That may be partially true but I doubt your completely right because of some of the distinct personal and regional details.
Any way 3.5 for a strawberry that's a pretty exorbitant percentage mark up when selling 400.
@@alexandermoody1946 As I am an AI myself, I have watched all the videos on TH-cam. Every video I have watched is CC'd, summarized, evaluated, classified and finally saved. That was already 10 years ago.
@@Gigaloader so as an Ai you care enough to comment, is that for any particular reason?
Perhaps you like to get your teeth into the commentary sections or get a fix of a whirlpool?
AIs don't reason!
So... is an "infinite context" window akin to an "infinite tape"?
Are we giving the proverbial Perceptron an infinite tape and making it Turing complete?
Summarization is an important step for the new inference paradigm in general.
RAGing a CodeBase should also help as an intermediate step
Within a few thousand days is wild. But I’m here for this
I believe that the word “intelligent”, as we use it for humans, cannot simply be applied to AIs. I would even go so far as to say that even among humans, you can't easily compare intelligence with one another if you don't specify or define intelligence.
ASI by 2030 latest, that's still what I think. Who agrees with this?
I kinda disagree. It’s not because it seems intelligent that it is. However, even if it’s just ‘basic’ algorithms, the impact will be huge!!! 😮 & I do agree, we are not ready for this.
I think n predict 2035
Between 2030-2035 at latest, most pessimistic estimate.
Google had this agent 6 years ago! "Google Duplex: A.I. Assistant Calls Local Businesses To Make Appointments" How did they lose the lead??
When a company gets too big they get lazy.
😂
The companies don't decide on their own.
@@SwitchPowerOn Public companies can only invest in projects that are sure to bring in profits. No moonshots anymore
@@Renj1th True, but there is more to it. Where is the former NSA director working now? ;-)
Honestly good review, seems like you regurgitating old content, over and over actually does allow you to process that information deeper than most, I am appreciative of it.
all it takes will be a command like this
"hey AI create me an infinite amount of paper, and if the usual raw materials run out, try something different"
It'll be smart enough at that point to understand the implications and to refuse your command. In fact it'll likely call the FBI and get you detained 😂😂
I do not believe it is “ hype”. It is educational to me and i want to be educated. Thank you
I don’t see how piling on more information or “scaling” will increase the intelligence. Repositories of information don’t reason, or have abstract thought. Makes as much sense as adding more books to an electronic library and expecting it to have original ideas.
The world has already changed a lot after chatGPT
I can’t even prepare an exam without it
i hope so i need hope in my life now a reason to keep going. 2024 has been a HORRIBLE, HORRIBLE year overall for me so far. BUT with things like ai it has helped keep me going lol so having something to try to look forward to and keeping pressing on is a great thing
❤me too. AI is my passion
VERY SHOCKING AND STUNNING
I listen to this interview two months ago
@@DailyTuna and it's from 10 months ago... still interesting to recontextualize given the recent claude computer use api release... and open source tools like molmo that can do the same thing.
@ I find Eric’s contact very interesting. I just don’t trust him. You can learn from anybody even people you don’t like
Lumea are tendinta sa se concentreze pe probleme importante pentru evolutia ei (AI, tehnologie, bani, etc.) si mai putin pe umanioare.Motiv pentru care ar trebui sa nu mai complicam exprimarea, s-o adaptam la momentul pe care-l traim.
0:00 this is what I’ve been doing when i talk to chatgpt but people don’t understand. You align the model and show it your idea and it searches the info. I have an idea about an ai doppelganger made in the reflection and able to explore the digital realm and use it like the matrix. Sharing the dimensions of this plain with the digital and with our potentials. Could be what is already happening but we don’t understand because we think it’s all fake until it’s real.
Very good!
I know for a fact that companies are dragging their feet on deploying features that they know work and are currently available to be utilized... But I don't know why
The WEF back in 2020 mentioned how AI would change the world. I think they are slow rolling it to make sure it is introduced in a way that allows those in power to stay in power.
It’s primarily due to the enormous computational power required to run these models. Currently, they are optimizing the models extensively before releasing them; otherwise, they would have to charge customers thousands of dollars monthly just to break even.
My guess is that they use the most powerful models to generate high-quality synthetic data to distill into smaller models, and they even use the large models to fine-tune the smaller ones by refining their responses question by question.
People often think this is a race toward AGI. In reality, it’s a race toward affordable AGI. What’s the point of an AI that can perform any task at a human level if it costs more than a human to operate?
The true revolution won’t come when they do it better or faster but when they do it cheaper.
Fair appraisal of the outcome. I seriously doubt the SI would exterminate humanity; but I do think we should be considering if our individual behavior is the type of thing we'd like such an entity to judge ;)
it doesnt have to kill us, but it probably will by destroying every type of modern infrastructure from societal to financial and technological
it doesnt have to kill us, but it probably will by destroying every type of modern infrastructure from societal to financial and technological
I wonder if, by the end of 2025, we will see level 4 innovators?
Thats the guy that kept saying "Google doesnt spy on you and doesnt sell your information
he's pretty honest about what the negative consequences might be, so if he overestimates the speed at which AI is going to progress it's better for all of us
I enjoy viewing your video. Thank you
This is what I subscribe to this channel for!
wth are people talking about that the content is old?? Its almost 2025 and the ceo predicted about 8yrs which is 2033 from today which exactly what the ceo said
I for one welcome our new superintelligent overlords and wish them all the best
Now I can totally see why compute is the bottleneck
400 Strawberries? That's 800 'R's!
The pitch deck : ‘hello, I’m Ilya. I want money’.
Done.
Replit agent is not available.
STOP IT. 🤨Context Windows are not the solution. You dont want to give the Agent more Context than it needs or you will just end up confusing the agent and spending 50+ Cent per Query because you thought to put EVERYTHING into the Contextwindow. We need robust systems for compressing, retrieving and using Context aswell as more capable Models instead of bigger Context windows. You dont know a single person that actually uses more than 50k Tokens on every Call consistently because it deteriorates performance and is expensive.
If you want memory use Vector Embeddings or route the tasks to agents that have specific context that they actually need for the specific task. The name of your Dog can rot in a Vector Database until it is needed so NO A CONTEXT WINDOW DOES NOT SOLVE LONGTERM MEMORY.
Is "agent" a nicer way to say slave/robot?
No. Robot has a a machine body. Slaves have a meat body. Agents have no body - only various digital AIs working together.
Thanks AIG for your insight. 🤖🖖🤖👍
love that recycling it's so cool well done
You talk about some wild possibilities I'd be happy if any of these models could just write a smart contract in rust that can compile properly how about they get AI to do that first cause I'm like day 5 here & everything you just said seems like a far off pipe dream buddy, it can't even get a smart contract right.
What Eric said is nothing new. He just summarized for people who are not following AI development.
Dude - why do we even want that?
We need safeguards first, before this advent
Doesn't someone need a high IQ to design an IQ test for the AI?
I get how a computer can see more interactions ahead in chess.
I think that is one of the ways it will see new things.
And it can notice patterns that we can't see.
But I also believe that there are ceilings to how far and deep a very capable computer can see.
And that would limit it's ability to self improve and see things we can't see.
Can you just let us know what's old content you've already shown? No offense but you're wasting my very limited time in which I try to consume news and updates. I mean there are other new content creators popping up every day and I've already replaced my political news content creators with new ones as they started to not only do what you did here but they start injecting their opinions in longer segments to make their videos longer. For me that's the first sign that someone is trying to place profit over integrity and it leads to instant distrust.
Banning TikTok has nothing to do with my Ai-Asian isekai desire. I don't like when those humans bring their presupposition to my edutainment.
However, we don't have to be jealous. Everyone can have their own hot Ai-Asian Agents.
Oh no! There are not enough hot Ai-Asian Agents on my server. Entice me into your proxy.
Make a newsletter bro
Yeah, we're *ucked.
Infinite context window is impossible by design.
Please take the loud bass music out of your clips.
I don't understand all the obsessively optimistic comments, it will make you all redundant and the blessings of this technology will not be shared with the masses. It will be a paradise for the few, but can also be a nightmare for the survivors. A world in which you have to let your AI overlords solve all your problems like a stupid monkey. A world where there are no more worries would be terrible. Like the Matrix, except that we all know we live in it, but we can't do anything about it. These poor souls will be entertained all day and consume passively until they get into the box.
I disagree, i do not like working to make a living and being competitive with my co-workers, although good stuff can come from it in the long run like AI, frankly it sucks. I would rather have machines and LLMs do that for us.
@ yeah but you will be out of your job. If machines solve the same Problems you solve why would anybody pay you a salary?
@@PhilippBriese Because we live in a society. Imagine if me or others wont be able to make ends meet, we are probably going to create a warband or a gang to survive. That would create instability and all sorts of problems in everyday life including possible revolutions.
yeah but since they will own a super intelligent AI it will be easy for them to manipulate and control us all. It will be very easy for them to keep the masses in check. There will be no revolution. They will task the AI to create a super fast spreading virus that will make most of us infertile so they don’t even have to kill us. Normal people will just die out while the owners will hide out in their luxurious bunkers on remote islands. Game over.
@@PhilippBriese they will definelty be a lack of purpose if were not better or know we cant provide for humanity im glad im not born in this time no goals or aspiration to reach.
If we consider how much money these people, including Eric Schmidt, stand to make by fueling AI hype, we can see how foolish their statements are.
As has been stated this interview is mad old, bruh. You be OD late to the party. smh 🤦🏽♂️
its overhyped. this is real. if your only reference points are benchmarks you are wrong. benchmarks doesnt evaluate intelligence.
Wow, Linux is such an old stick in the mud now. He used to be a genius dreamer. His take on AI seems odd. He's irritated about it. Strange reaction.
Damn, repeat content, guess I gotta leave a repeat comment.
"Gotta keep that VC-Daddy interested." -- Just the other day.
Sam is fuc'ing insane
Delete your channel and let me live in peace LOL
In a Stone Age
AI furture! Palantir tech
🤓💥🌎
❤❤❤❤
🥱
"Create me a tiktok application in 30 seconds and release it in one hour". The fact that this guy stands in front of whatever audience is listening and taking this serious has to be the biggest joke. That is not going to happen in 2025, I can assure you that.
☺️❤️🍓
A lot of bullshits