I feel like you missed the point of the Turing test. It is not about really testing whether a system is conscious. It is a thought experiment showing us our limitations to verify consciousness. It is the same as if I asked you to convince me that you are alive. I cannot see your inner workings, your thoughts or "soul" (whatever you want to call that thing that distinguishes us from dead matter). So the only way to convince me is through the error prone process of communication. If a "dead" pattern-predicting system was trained good enough to convince me that it is alive, I am just not able to distinguish it from a system that really is alive when using the same channel of communication.
Not the best explanation of current machine learning models. Yes, they will make a prediction with a given percentage of certainty, but they don't necessarily just parrot back a response they've seen before as you suggested multiple times in this video. Wholly original sentences can and often are created. I think this piece was very poorly explained.
This video was much more convincing than I'd expected it to be. I've never heard someone make this point before and you made a really compelling argument.
Brother don't take it the other way but being a cs engineer I would recommend you to research the concept of AI/ML a lil bit more and also to read a subject called TOC (theory of Computation or automata theory) if you wanna know about Turing machines and etc., Some of your concepts are very wrong from the very base of it so no matter how much effort you have given on researching the fancy or attractive part of it.........it sounds baseless. Hope you understand
"It's not actually intelligence it's just pattern recognition" is one of those sentences that make me wonder how much we over estimate our own capabilities.
This is false…. It doesn’t “look” at billions of answers to previous questions it only feeds the question into a mapping that has already been fine tuned against those billions of previous relationships. The resulting answer is then the best guess based on the relationship devised from the training session of the billions of examples.
alan turning predicted mobile phones already “One day ladies will take their computers for walks in the park and tell each other, "My little computer said such a funny thing this morning”
Great video as always. Just pointing out though that his death was in 1954, not his birth (1912). His efforts breaking the Enigma code were widely regarded as one of the single biggest achievements of WW2 by a single person, he later died by suicide.
AI aren't search algorithm, they don't search for the closest answer, but use a general understanding derived from their training to craft a complete output matching the input.
I realise I accidentally said ”born in 1954” instead of ”died in 1954”
died in 1954
I feel like you missed the point of the Turing test. It is not about really testing whether a system is conscious. It is a thought experiment showing us our limitations to verify consciousness. It is the same as if I asked you to convince me that you are alive. I cannot see your inner workings, your thoughts or "soul" (whatever you want to call that thing that distinguishes us from dead matter). So the only way to convince me is through the error prone process of communication. If a "dead" pattern-predicting system was trained good enough to convince me that it is alive, I am just not able to distinguish it from a system that really is alive when using the same channel of communication.
Not the best explanation of current machine learning models. Yes, they will make a prediction with a given percentage of certainty, but they don't necessarily just parrot back a response they've seen before as you suggested multiple times in this video. Wholly original sentences can and often are created. I think this piece was very poorly explained.
"I am not fearful of the day an AI will pass the Turing Test, I am afraid of the day it fails intentionally." Is this the beginning of the end?
Did you even say the word google one time in the whole video? Your title is misleading.
This video was much more convincing than I'd expected it to be. I've never heard someone make this point before and you made a really compelling argument.
the majority of humans cannot pass the turing test at all.
I respect ❤️you but over the time kalle ur video is less informative😕😕
Brother don't take it the other way but being a cs engineer I would recommend you to research the concept of AI/ML a lil bit more and also to read a subject called TOC (theory of Computation or automata theory) if you wanna know about Turing machines and etc., Some of your concepts are very wrong from the very base of it so no matter how much effort you have given on researching the fancy or attractive part of it.........it sounds baseless. Hope you understand
Born 1912, Died 1954
"It's not actually intelligence it's just pattern recognition" is one of those sentences that make me wonder how much we over estimate our own capabilities.
This is false…. It doesn’t “look” at billions of answers to previous questions it only feeds the question into a mapping that has already been fine tuned against those billions of previous relationships. The resulting answer is then the best guess based on the relationship devised from the training session of the billions of examples.
Exactly what I'm being saying to my friends after the google thing got public. Finally someone in the same line of thought. Cheers.
alan turning predicted mobile phones already “One day ladies will take their computers for walks in the park and tell each other, "My little computer said such a funny thing this morning”
Great video as always. Just pointing out though that his death was in 1954, not his birth (1912). His efforts breaking the Enigma code were widely regarded as one of the single biggest achievements of WW2 by a single person, he later died by suicide.
Is there something wrong with me if I predicted the tenth number to be 0 ? :)
AI aren't search algorithm, they don't search for the closest answer, but use a general understanding derived from their training to craft a complete output matching the input.
I am surprised companies don't use their recorded phonecalls to train ai, people talk very differently online compared to in person
Nice video! Very interesting.