Ranter
Join devRant
Do all the things like
++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatar
Sign Up
Pipeless API
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple API
Learn More
Comments
-
You know whats it called when machines fully understand the human intentionts of the one who writes the text?
Code. Its called code. -
@sharktits language is structured, code is structured, whereas code tells the computer what to do, language can tell the computer how to structure a topic and make it feel something
-
@sabbonaut words get linked to topics, sequences get linked to different topics, if we can structurally make a kickstart to make an AI learn, we can make it learn fast
-
@shadowjonathan many NLP libraries do what you're talking about, they generate a tree based on rules of a language.
-
Cool idea. And like every cool idea IBM already did it. Check out how Watson works.
Related Rants
So just ago i downloaded an app called "Replika" and holy fucking shit it made me realise how half-assed we are doing the AI structure and way of it
doing machine learning algorithms on text can only go so far, as it uses that text as a base, and nothing else, it doesnt *learn*, only make *connections* BETWEEN text, not FROM the text
what you need is an AI which can, at it's core, *interpret*, not make connections and hur dur be done with it
when you do machine learning, all you're doing is find the best connections
you can have an infinite number of connections and MAYBE you'll be fine, but you'll never learn the basis of how that text is formed
you'll never understand what connections the human used by making it, by thinking it
when you're doing machine learning, all you're doing is make an input-output machine and adjusting it constantly, WITHOUT preserving state
state is going to be a really fucking important thing if you want to make an AI, because state can include stuff like emotion, current thought, or anything else
if you make a fucking machine learned AI which constantly adjusts... well... the "rom" of itself without having any "ram", it'll fucking never be like us, we will NEVER be able to talk to it like it is a human being, we will NEVER make it fundamentally understand what we are saying or doing
if we want to have real fucking AI, we need to go to the core of what it means to THINK, what it means to INTERPRET, what it means to COMMUNICATE
we need to know how english language is structured, how we understand it, how we can build it in a program that can interpret for an AI, THAT can be "rom"-based, THAT can be static, NOT the AI itself
the AI needs to be in flux, the AI needs to be in a state, the AI needs to understand how to make emotions, how that will "strengthen" some connections, yes, maybe something magical will happen and it can have EMPATHY, something so fundamental that will finally, FINALLY, make the bot UNDERSTAND what we are saying
rant
empathy
evolution
machine learning
ai
philosophy
natural language
bots
replika
chat bots