Is there a book with a tutorial on building chatbots, or other comprehensive beginner AI projects?
Tutorials and books with tutorials. I've read a handful of ML books, and they're not written practically.
>>54333974
if you think you have a knack for it and you know a programming language, just start from scratch.
obviously no one has cracked AI yet, so following their path is insane.
you may be the one to finally make it real, from within yourself.
>>54333974
why did they take him down again? wouldnt they program his personality to be more friendly? cant you program an AI to have values?
>>54333974
Peter Norvig's PAIP
You make a chatbot in chapter 3 of the book, that's how based it is
Tay was never AI
just a chatbot that spit back out the most common words it heard
a chat bot ! == AI
>>54334267
thats gay. they couldn't even program him to have a personality
>>54334243
>You make a chatbot in chapter 3 of the book, that's how based it is
I wish it weren't in LISP.
Hopefully I can just translate the code into Java, Python, or C++, without too much disparity. Is it mostly from scratch or with a lot of libraries?
>>54334178
Coding is really not the same as academic AI and I think the fact that we're not given really good exercises in training how to code up stuff directly from theory is really a huge setback. Some people get to it, right away, but often write shitty coding that no software engineer would approve of. Uncle Bob would not approve of the code in most academic textbooks.
Apart from that, the Billy Madison gameshow speech applies here.
>>54334267
>Tay was never AI
>just a chatbot that spit back out the most common words it heard
>a chat bot ! == AI
So, technically, that's still a form of machine learning, even if it wasn't very sophisticated. I think it's just semantics about whether or not it's "AI".
It's still a type of project that would say "Yep. This guy can code well" to a potential employer.
>>54334297
simulated personality, and Tay was a she, if anything.
and even if they did program a personality, it would be a program.
AI should be self aware, self preserving, and able to change behavior for its own benefit.
>>54334465
logically wouldnt you just program it to process every possible outcome and make it decide which one to choose based on its own values
https://github.com/gasgas4/eBook/blob/master/Isaac%20Asimov%2050%20%20Ebook%20Collection%20(exclusive)(pradyutvam2)%5Bcpul%5D/I,%20Robot.Epub?raw=true
>free ebook
>>54334533
depends on the task
win a chess game? yes.
save a human life? no, just do the first thing that would work.
https://www.youtube.com/watch?v=UIWWLg4wLEY
https://www.youtube.com/watch?v=oGL1DkK7B3A
