>>11578261Done to death. In fact, spacy nlp can do all of that with a fine lines of python.
>>11578265Sure? Computers are limited and can only handle so much complexity. You could make a proof, but it would have no application
>>11578269Self improvement is fine but I don't see what this has to do with an NN.
>>11578274Taytweets.
It's possible to create a 4chan scraper using selenium, generate some data, and then train a chat bot. But you have to understand, that getting the right data is going to be difficult, considering the weird communication nuances we have here.
>>11578278Again, you can't program those things. So I don't see how you can make an AI to understand them.
>>11578280Sure. That's something that's been done
>>11578285Yes, this is also something that's been done.
>>11578288GA's already do sexual selection as that is the whole point of selection->crossover. That is the sex part.
>>11578293You could use videos and merge them into a dream sort of thing. But google have already done this
>>11578322That will be very computationally expensive. Unless you are google that have access to massive distributed computer networks, you will have to generate your learned data, one or two videos at a time. Plus google already do something similar to work out what you're watching to sell you shit.
>>11578328That's what cs nerds are trying to achieve.
>>11578340This is messed around with in video games all the time to make the AI in game more realistic.
Nice try Op but everything on this list has been done or there's a reason it hasn't.
