22 y/o NEET with no programming experience besides basic LUA scripting on Roblox [Stack Exchange style hehe] >inb4 /pgt/ I want into AI programming, is this a good roadmap github.com/AMAI-GmbH/AI-Expert-Roadmap
Basically I want into GPT-3 but better. Seems like that path would be depending on OpenAI's efforts which doesn't seem ideal for a Jarvis type AI.
Also from my understanding I would need to shell out 10k to Google's TPU v4 which is shown to outperform Nvidias A100.
The problem I find with this I won't be able to go offline as I'd be depending on Google's uptime, which may be seen having outages [my predictions due to Azure going down quite frequently since the pandemic began]
I think pretty far as business majors don't even take real calculus.
Elijah Long
I know that but was just wondering how big a shock the difference will be , I guess the question is self answerable
Lucas Butler
Just look up edgy hashtags on Tiktok. Boom.
Joshua Gomez
no one cares about AI in the real world
Joseph Perry
That is a scene girl not an emo girl.
Gabriel James
I don't agree with that graph but whatever.
you don't need a fucking dedicated TPU to learn AI. You need something like that if you're getting paid and you're training production grade models.
Also you won't be building Jarvis anytime soon, if you're 22 now then maybe by the time you're 55 there may be something somewhat similar to that but I doubt it. And things like that aren't even the result of a single person, at most you would be in a team
Personally I recommend doing statquest in order. By the time you're done with it you will have made a few AIs and you'll have a nice understanding of them. Then you can dive deeper into the math to learn more. The reason why I don't like that roadmap is that it expects you to learn the equivalent of 2 degrees before you even start working on AI. That's ridiculous you can get started working on them in a week or two. There's no reason to delay getting your hands dirty. All that formal, academic knowledge is kinda necessary to get a degree and to get a job, but seriously academia is way overrated. You can do the same work a data scientist does is you do all of statquest
Matthew Price
that roadmap is for becoming an academic researcher with a phd
Angel Johnson
Except without a PhD because self studying doesn't give you the certificate that will get you through the door