Juan also talks about AI that's being used against us (probably Twitter accounts) around the 20min mark. You guys should check out the gpt-3 ai. Lots of implementations, where they can make automated hate speech, translate nasty speech into friendly speech that has the same meaning and sounds like a person, it's crazy. I wouldn't put it past them to have similar AI making doomer posts here.
There is something to understand, seems that Juan is not a computer savvy , there are AIs that you can train to understand writing and emotions in writing etc... but they are not sentient, AI are very complex algorithms, they still rely on huge set of data and a very good software... The software doesn't change what changes is the data association.
Search for NLP (Natural Language Processing)
Now, you tie that with a really big computer and a really big database and time to train the algorithm and you have an interesting tool that can analyze a wide range of information and give you interesting results.
Remember Captcha that was asking to recognize some pictures to know if you are a human? Well that's a tool to train a computer AI, it doesn't "learn" per se, it's trained in the sense that when it asks about a Bicycle and you show him what a bicycle is, its data set grows a lot, and it's trained by a myriad of humans for free!!
In summary, AI are not intelligent as in humans, they have a big dataset and enough computational power to answer fast.
Juan also talks about AI that's being used against us (probably Twitter accounts) around the 20min mark. You guys should check out the gpt-3 ai. Lots of implementations, where they can make automated hate speech, translate nasty speech into friendly speech that has the same meaning and sounds like a person, it's crazy. I wouldn't put it past them to have similar AI making doomer posts here.
There is something to understand, seems that Juan is not a computer savvy , there are AIs that you can train to understand writing and emotions in writing etc... but they are not sentient, AI are very complex algorithms, they still rely on huge set of data and a very good software... The software doesn't change what changes is the data association.
Search for NLP (Natural Language Processing)
Now, you tie that with a really big computer and a really big database and time to train the algorithm and you have an interesting tool that can analyze a wide range of information and give you interesting results.
Remember Captcha that was asking to recognize some pictures to know if you are a human? Well that's a tool to train a computer AI, it doesn't "learn" per se, it's trained in the sense that when it asks about a Bicycle and you show him what a bicycle is, its data set grows a lot, and it's trained by a myriad of humans for free!!
In summary, AI are not intelligent as in humans, they have a big dataset and enough computational power to answer fast.
I understand your sentiment, but check this out, a conversation with a person from a gpt-3 implementation:
https://lacker.io/ai/2020/07/06/giving-gpt-3-a-turing-test.html