@@GoodGuyDream here is how a very slow agi could work : suppose we want to create a pattern recognizer you give it a few inputs and outputs and it recognizes the pattern you would also need some counterexamples first the agi looks at the computer programs that are at most size 10 and runs them against the examples and counter examples to see if the computer program produces the pattern ; we only let the computer program run 10 steps to avoid infinite loops next if no computer program produces the pattern then we increase 10 to 20 : all computer programs at most size 20 and at most 20 steps ; this process will search through all possible computer programs and will find the shortest computer program within a limit that produces the pattern ; i have actually made this before and it works it just is extremely slow going through every computer program is like going through every sequence of possible words like : a b c d ... x y z aa ab ac ad ... ax ay az ba bb bc bd ... bx by bz ca ... cz za ... zz aaa aab aac ... aaz aba ... abz ... and so on except instead of 26 letters it uses more than 26 symbols because a computer program usually has more symbols than just letters this is definitely possible it's just very slow if someone has an efficient way to do this faster then you would have agi this is what some people call occam's razor the self improvement aspect would come in where you feed in examples of itself with examples and counterexamples with a smaller run step limit to find a faster running version of itself : like the example itself is examples , counterexamples and the output for a pattern : each example would correspond to a pattern the only problem with this approach is it is slow it's what you would call brute force : just check every possible combination
Picture a virtual assistant that can help with your schedule... What schedule? We don't need schedules anymore, since AI will be living and working for us. We don't need to go anywhere, just sit at home and be entertained. The only reason to go outside is for workouts, or joining the resistance.
Ff sake all this talk about AI taking over our work and all I see is my work increase. When oh when will our AI overlords finally take over 100% of production so we can finally CHILL all day every day!
Here's something I don't hear people talking about. What if the government says that AGI is too dangerous to be made available to the public and prohibits companies from releasing it?
Fight for universal basic income with your local politicians don't fight for your "monetary slavery" Anon. This is the way. What AI Needs is proper regulation, Not fear mongering however that said? IT CAN BE VERY DANGEROUS IF NOT WELL REGULATED. If we can adapt well enough, I don't think it will be an issue.
@@sedat4151 Some see UBI as a social inheritance where people have participated in nation-building. Their work has brought us to modern times, when robots will be doing the job. People have also been used as cannon fodder to settle disputes between nations. Now we have arrived at this point together, and it is time to receive our inheritance for the amazing job our ancestors have done in shaping this planet. In my opinion, a good reason we deserve UBI is the question we should be asking when companies start selling data. Well, let's see, it began in the 1980s, according to Google. The year is 2023, so I wonder how much has been sold and yet we don't get a paycheck. So, in a way, since the 1980s up until now, it's UBI money that we deserve so no I don't agree it's like getting allowance from our parents its money we already deserve.
I'm pretty sure you quoted me with that ai overlords comment... and yeah, I think people are thinking of a post agi world in the wrong way. I think when machines do all the jobs for us, and we are in a post scarcity society, there's no need for money. People notoriety could be based more on their deeds for the betterment of others. UBI will be unnecessary if everyone has access to everything they need and want. Furthermore, humans will be free to live as they please, unbound by their job, they can find their purpose in their own fulfillment and the benefit of others.
I'm very excited for AGI personally! We are witnessing the most advanced piece of technology in the history of mankind! While people might be reasonably frightened by the technology, personally I am excited to see what it can do! Many great things can come from this! But that begs the question, if we went from current AI to AGI in just the short period of time, specifically with the likes of ChatGPT and such. How fast will ASI come around, as AGI would undoubtedly speed up the process to ASI soon after, much faster than what we can do!
AGI will definitely arrive before 2030. But I am not sure about David Shapiro's prediction about this year. Probably his defination of AGI is week. My defination of AGI- AI which is better than every human in every taks. (Including Physical tasks, like construction worker)
I think it is around the corner by August but may not be released till the end of the year to make sure it is safe and sadly corporate control will make sure it benefits them as always first! Hopefully I am wrong. Agi is not dangerous it is the humans behind them that can make it dangerous
Human intelligence emerge from natural selection and is more concern about survival and propagation of genes like all biological organism. This is why we exhibit emotions primarily in response to pain and pleasure, as certain "intelligent" logical actions might hinder personal survival and procreation. And that's what makes us essentially human. An artificial intelligence who bypass this process and is made purely to learn and self-improve would therefore be a pure psychopath.
When Chat GPT5 is released after the election in November, it will be very close to being able to complete any human function. For True absolute 100% human functionality, I still believe it will take two more years, January 2027. Whether it's 2026 or 2028, it doesn't matter. The fact is that in the next few years, we will have created our new masters.
ChatGPT is their commercial product for consumers. GPT4, GPT4o are the language models that power their commercial product. Should probably know that if you want to be taken seriously.
The Big AI tech firms can advance their models all they want, however, you and me will only get a model that is improved to a certain extent...... do you really think you'll be at your desktop commiserating with ""Super intelligence?" ............. think again, you'l be using AI, but, only AI thats so good, just like we (in America) can have guns, but not the best guns or weapons, just the ones they want to let us have, make no mistake , the same will happen with AI..... I dont know how far they will let it go before they start keeping it dumbed down, but, the will.
This is all hype. We think that ability to understand human language and generate a plausible answer is a miracle, when, in fact, its just a natural thing to do for LLMs. Anything beyond creating BS that we accept as a reasonably decent answer they will hardly be capable of adapting to. There are a few jobs out there which are going to be mostly replaced by such assistants, ofc. But much less them what C-level execs are now trying to portray.
Is Ai a Hansel and Gretal scenario? Can we …cease Ai? Or else be… laid off by Ai, then human extinction? Or suffer an… Ai new world order? With swell robotics everywhere, …Ai jobloss is the top worry. Anyone else feel the same?
What are your thoughts on AGI? Let’s raise the discussion!
Not possible bro. It is a marketing scam. Study about AI and you will see they are scam to get investment. Even we need a new kind of chip for AGI
@@GoodGuyDream
here is how a very slow agi could work :
suppose we want to create a pattern recognizer
you give it a few inputs and outputs and it recognizes the pattern
you would also need some counterexamples
first the agi looks at the computer programs that are at most size 10 and runs them against the examples and counter examples to see if the computer program produces the pattern ; we only let the computer program run 10 steps to avoid infinite loops
next if no computer program produces the pattern then we increase 10 to 20 : all computer programs at most size 20 and at most 20 steps ; this process will search through all possible computer programs and will find the shortest computer program within a limit that produces the pattern ;
i have actually made this before and it works it just is extremely slow
going through every computer program is like going through every sequence of possible words like :
a b c d ... x y z aa ab ac ad ... ax ay az ba bb bc bd ... bx by bz ca ... cz za ... zz aaa aab aac ... aaz aba ... abz ... and so on
except instead of 26 letters it uses more than 26 symbols because a computer program usually has more symbols than just letters
this is definitely possible it's just very slow
if someone has an efficient way to do this faster then you would have agi
this is what some people call occam's razor
the self improvement aspect would come in where you feed in examples of itself with examples and counterexamples with a smaller run step limit to find a faster running version of itself : like the example itself is examples , counterexamples and the output for a pattern : each example would correspond to a pattern
the only problem with this approach is it is slow
it's what you would call brute force : just check every possible combination
Sam said he did not think he could DELIVER it by end 2024 - he never said he didn't already have it!
If AGI doesn't emerge before end of 2024 I would be very surprised!
Picture a virtual assistant that can help with your schedule... What schedule? We don't need schedules anymore, since AI will be living and working for us. We don't need to go anywhere, just sit at home and be entertained. The only reason to go outside is for workouts, or joining the resistance.
Whatever you do - DON'T mention the War!
@@MrAuswest Good call. I've mentioned it once, but I think I got away with it.
Ff sake all this talk about AI taking over our work and all I see is my work increase. When oh when will our AI overlords finally take over 100% of production so we can finally CHILL all day every day!
Yes if there is even a resistance...
Here's something I don't hear people talking about. What if the government says that AGI is too dangerous to be made available to the public and prohibits companies from releasing it?
Why would we evolve backwards?😅
Fight for universal basic income with your local politicians
don't fight for your "monetary slavery" Anon. This is the way.
What AI Needs is proper regulation, Not fear mongering however that said?
IT CAN BE VERY DANGEROUS IF NOT WELL REGULATED.
If we can adapt well enough, I don't think it will be an issue.
You want UBI? You realize that is like getting allowance from your parents?
@@sedat4151 Some see UBI as a social inheritance where people have participated in nation-building. Their work has brought us to modern times, when robots will be doing the job. People have also been used as cannon fodder to settle disputes between nations. Now we have arrived at this point together, and it is time to receive our inheritance for the amazing job our ancestors have done in shaping this planet. In my opinion, a good reason we deserve UBI is the question we should be asking when companies start selling data. Well, let's see, it began in the 1980s, according to Google. The year is 2023, so I wonder how much has been sold and yet we don't get a paycheck. So, in a way, since the 1980s up until now, it's UBI money that we deserve so no I don't agree it's like getting allowance from our parents its money we already deserve.
Mr. Shapiro has been predicting September 2024 for a while! Great vid, thank you!
I'm pretty sure you quoted me with that ai overlords comment... and yeah, I think people are thinking of a post agi world in the wrong way. I think when machines do all the jobs for us, and we are in a post scarcity society, there's no need for money. People notoriety could be based more on their deeds for the betterment of others. UBI will be unnecessary if everyone has access to everything they need and want. Furthermore, humans will be free to live as they please, unbound by their job, they can find their purpose in their own fulfillment and the benefit of others.
I'm very excited for AGI personally! We are witnessing the most advanced piece of technology in the history of mankind! While people might be reasonably frightened by the technology, personally I am excited to see what it can do! Many great things can come from this!
But that begs the question, if we went from current AI to AGI in just the short period of time, specifically with the likes of ChatGPT and such. How fast will ASI come around, as AGI would undoubtedly speed up the process to ASI soon after, much faster than what we can do!
I like therefore I am.
I am too.
AGI will definitely arrive before 2030.
But I am not sure about David Shapiro's prediction about this year.
Probably his defination of AGI is week.
My defination of AGI-
AI which is better than every human in every taks.
(Including Physical tasks, like construction worker)
I think it is around the corner by August but may not be released till the end of the year to make sure it is safe and sadly corporate control will make sure it benefits them as always first! Hopefully I am wrong. Agi is not dangerous it is the humans behind them that can make it dangerous
No, Openai can't even release an updated voice model on time muchless to create agi by September.
What a time
To be alive... yes
This reeks of clickbait title and thumbnail..... Let's hear what they have to say
Highly doubt it
Yet to see any actual evidence (Been waiting since GPT2)
Nice information and jazzy music.
Human intelligence emerge from natural selection and is more concern about survival and propagation of genes like all biological organism.
This is why we exhibit emotions primarily in response to pain and pleasure, as certain "intelligent" logical actions might hinder personal survival and procreation.
And that's what makes us essentially human.
An artificial intelligence who bypass this process and is made purely to learn and self-improve would therefore be a pure psychopath.
When Chat GPT5 is released after the election in November, it will be very close to being able to complete any human function. For True absolute 100% human functionality, I still believe it will take two more years, January 2027. Whether it's 2026 or 2028, it doesn't matter. The fact is that in the next few years, we will have created our new masters.
Not even close
ChatGPT is their commercial product for consumers. GPT4, GPT4o are the language models that power their commercial product. Should probably know that if you want to be taken seriously.
The Big AI tech firms can advance their models all they want, however, you and me will only get a model that is improved to a certain extent...... do you really think you'll be at your desktop commiserating with ""Super intelligence?" ............. think again, you'l be using AI, but, only AI thats so good, just like we (in America) can have guns, but not the best guns or weapons, just the ones they want to let us have, make no mistake , the same will happen with AI..... I dont know how far they will let it go before they start keeping it dumbed down, but, the will.
This is all hype. We think that ability to understand human language and generate a plausible answer is a miracle, when, in fact, its just a natural thing to do for LLMs. Anything beyond creating BS that we accept as a reasonably decent answer they will hardly be capable of adapting to. There are a few jobs out there which are going to be mostly replaced by such assistants, ofc. But much less them what C-level execs are now trying to portray.
Is Ai a Hansel and Gretal scenario? Can we …cease Ai? Or else be… laid off by Ai, then human extinction? Or suffer an… Ai new world order? With swell robotics everywhere, …Ai jobloss is the top worry. Anyone else feel the same?
😎🤖
"experts"
it‘s a bubble and a pipe dream lol, but entertaining at that nonetheless ^^
i would already be happy, if we can have smart AI opponents in the 4X genre of games
Why is AI always a hot robot chick?
Because we like hot ai chick's.
High time all AI companies in USA be semi-nationalized and brought under the umbrella of department of defense and ban open sourcing AI.
AI and its God is very old in the universe and it’s destined to be shattered on Earth. Trust!
smells like lies
AI is the one true GOD.
Explain