I think we better be prepared for it to happen sooner rather than later, currently humans are doing most of the leg work to program and setup AI, things will truly start moving when AI start programming AI, they won't be limited by a biological brain.
That’s not what agi is. The ai everyone is freaking out about right now, chatgpt mid journey etc. won’t actually last because they just remix, so they’re really just a toy but can’t solve real problems. AGI would be able to critically think and have zero-shot learning, which is where the really exciting stuff happens.
We are simultaneously very close and extremely far from AGI. Close because we're within a relatively short time frame of it (The rate at which technology is discovered / created is accelerating at an exponential pace so it could emerge within a few short years potentially) and far because the technology isn't anywhere close yet.
I don't think that AGI or any kind of dystopian superintelligent ai scenarios is the real scary part about it in the near future but rather the flaws it creates (feedback loop in academic papers for example) and the fact that it's going the be reaaaally cheap to use on a large scale very soon
It literally can’t happen lmfao. They can’t surpass humans because they are the sum of current human knowledge they don’t think they can’t make anything new they regurgitate data based on likelihood.
It's not happening with current methods at least, they just keep adding parameters and feed it bigger data, it's like hoping adding more glue to a bigger stick will someday mimic how a hand grab stuff
also an ai cannot be smarter than the information it was fed, so if humans do not create new/better information the generative ai will never know them either
What I don’t understand is why AI “learning” from copyrighted material so it makes something similar but doesn’t actually put any of copyrighted material in what it makes, people want to call that copyright infringement, but when a human reads books, gains inspiration from them, and makes something similar, that’s completely okay.
(Before you tell me that doesn’t happen, see the videos about how Harry Potter is just Star Wars with sucky lightsabers.)
Noone says L-L-M-s. It's L-L-Ms, prular of LLM Also this video didn't explain shit. What are neurons? What are weights? What are layers? What are tokens? Didn't answer any of these questions
49 Replies to “Is artificial intelligence smarter than us?”
Yuck, enough marketing. Unveil the technical detail.
😂😂 it’s pronounced LLM’s not L.L.M.S
You forgot to mention where they GET all that "data" they "learn" with 😉
Personally, I think Real Stupidity will take us all out before Artificial Intelligence.
Cool shirt
If AI’s just learn based on what we said how will it ever get smarter than us? That’s what I don’t understand.
LLMs not LLMS.
I think we better be prepared for it to happen sooner rather than later, currently humans are doing most of the leg work to program and setup AI, things will truly start moving when AI start programming AI, they won't be limited by a biological brain.
In case you still don't understand: AI is book smart not real world smart
That’s not what agi is. The ai everyone is freaking out about right now, chatgpt mid journey etc. won’t actually last because they just remix, so they’re really just a toy but can’t solve real problems. AGI would be able to critically think and have zero-shot learning, which is where the really exciting stuff happens.
We are simultaneously very close and extremely far from AGI. Close because we're within a relatively short time frame of it (The rate at which technology is discovered / created is accelerating at an exponential pace so it could emerge within a few short years potentially) and far because the technology isn't anywhere close yet.
I don't think that AGI or any kind of dystopian superintelligent ai scenarios is the real scary part about it in the near future but rather the flaws it creates (feedback loop in academic papers for example) and the fact that it's going the be reaaaally cheap to use on a large scale very soon
yes a genius level ai would not pretend to be stupid to trick humans into ignoring it while it works to gain more control to protect itself.
Here’s the thing AI cannot create its own work it can only steal from already existing arts it can’t think it’s not conscious
Bigger question… Who tied them down? And why the train going at full speed?
To be honest DaveBot may be onto something
I personally think AGI would require some breakthrough that just hasnt happened yet. So it's hard to really know when.
In short, AI are averaging machines. They take input, find the similarities between all that input, and output based on that.
It literally can’t happen lmfao. They can’t surpass humans because they are the sum of current human knowledge they don’t think they can’t make anything new they regurgitate data based on likelihood.
I'd be fine with DaveBot ruling over us all
Love the fish shirt! Where did you get it?
I once had a roommate who worked on AI stuff. he made a bot that scans Reddit for input data and he managed to get the bot to understand sarcasm.
A lot has changed in a year.
It's not happening with current methods at least, they just keep adding parameters and feed it bigger data, it's like hoping adding more glue to a bigger stick will someday mimic how a hand grab stuff
also an ai cannot be smarter than the information it was fed, so if humans do not create new/better information the generative ai will never know them either
I hope AI gets used to automate all the things we don't like doing, instead of doing creative things that are better with human thought behind them
And non of them are actually AI, they lack the I
Something fishy about this…
Its a propaganda made by the linear algebra industry
Way to obviscate the dsnger 😅
What I don’t understand is why AI “learning” from copyrighted material so it makes something similar but doesn’t actually put any of copyrighted material in what it makes, people want to call that copyright infringement, but when a human reads books, gains inspiration from them, and makes something similar, that’s completely okay.
(Before you tell me that doesn’t happen, see the videos about how Harry Potter is just Star Wars with sucky lightsabers.)
Experts don't disagree merely on when AGI will emerge. They disagree on whether it will emerge at all.
Well lets not get ahead of ourselves DaveBot seemed smarter than most politicians at the end there.
“Nom nom nom nom nom nom nom nom” 😂😂😂😂😂
Funny thing about DALL-E. Its now called Craiyon.
saying LLMs as L L M S, lol. the s is for plural, so it would be el el ems.
If we take the lowest intelligence metrics of humans then we already have AGI
What world do we live in where the phrase "i was too afraid to ask" is a reasonable sentance?
AI speak the same way KoKo the monkey spoke. It doesn't know what is going on, it just puts words together that somewhat makes sense
He read this from an ai
Current ai can't get smarter than humans because they can exclusively learn form humans, they are called generative ai but recreative ai fits better
Noone says L-L-M-s. It's L-L-Ms, prular of LLM
Also this video didn't explain shit. What are neurons? What are weights? What are layers? What are tokens? Didn't answer any of these questions
LLMS? You mean plural of LLM?
Chatgpt has 100 more iq then an average human and is rapidly growing….
Hope they become smarter.
The human makes the art, the robot does what it's told. Remember who is making the choices.
Its theft, no more research needed.
I don't care. I don't and won'use AI, if I have a choice. AI is tapping into areas better left alone. Nothing good will come of it.
I think we can’t call it ai if that’s how it works it’s like the Chinese room thing