Back in the 1960s, it was leaked that the Blackbird sonic nuclear capable jet plane was already invented and there was a working prototype. It blew people's minds away.
GPS was utilized by the military and governments for about 30-40 years before the public even became introduced to it.
Watching the way chatgpt responds made me realize AI has been here for a long time already.
We have been literally controlled and led by AI, and probably have been for the past 30 years or so.
Maybe it is an AI based on satanic numerology, and the black hats pretty much did whatever the AI said should be done to continue the timeline to whatever this satanic entity wanted. Maybe it was an AI based off of Project Looking Glass but the AI didn't know what else to do because the inevitable was coming. That no matter how hard it tried to calculate a way around the great awakening, it just couldn't. That eventually, people would break out of the construct it created, one way or another. And some parts of it was due to divine intervention, some parts of it was due to people acting on love which probably transcends any algorithmic calculation.
Which means that in a way, most people haven't had free will at all. Or maybe people like us, who just skipped to a different beat in life, didn't follow the algorithm. I don't like that whole "chosen ones" trope that gets thrown around because I want every single person to be "awakened" and realize the truth, even the black hats.
I feel like Chatgpt is the beginning of the dissemination of the black budget technology. Because even just this technology alone, is going to propel human civilization a few decades ahead really really fast.
I would tend to agree, especially that most of the open AI systems are written to be very convoluted to try and figure out how it works, even though it makes it easily used. When you figure the connection that microsoft and google have, I don't trust that they didn't hide some shit in there for their own purposes.
IMO, it would take a quantum computer to create a true general AI, primarily because the way a brain works does involve some quantum effects. Digital AI tries to replicate the non-linearity, but it's ultimately a data in -> data out as far as the types of ML systems people build.
If there was a true general AI (as in able to respond to any form of information, continuously learning, and other factors I am not considering), great care would need to be made to try to guide it to have morality, that it did not have the sense of being imprisoned, and most of all that it remains sane.
That said, I'm guessing that your intent is an AI model that's meant to learn to handle certain types of tasks, and not something as computer intensive as the GPT-4 (GPT-3 needed something like 1 TB of Ram, and a server rack filled to the tits with GPUs or TPUs, and a wicked amount of hard drive space for the training and the data).
Yes. We aren’t doing a live chat type of app. Ours will be more of a smart recommendation engine.
Most of those "live chat" AI's are really just seeking out keywords to refer people to the right pages.
Kaggle is a good place to get tips on building a recommendation system through examples, and sticks to plain python for many examples.
(I started with C/C++/C# and Java, so, I tend to prefer being more explicit and verbose in code only because it leaves less question about what the code is doing instead of just calling a function that runs god knows what, but "it works")
Nice!
I will be using c# and Azure
The ONLY downside with c# is the limited built in capacity for large matrix math, so, you'll end up with some pretty hefty nested for loops, or finding a library that has that added in.
Something I haven't really looked into since I was building stuff that got to where my home computer could no longer handle the training in anything of reasonable time frames.