The problem with AI and teaching it (very minor problem), is there are now so many layers to it. All of them are pretty teachable. But when someone says, I want to learn AI, there is an ambiguity in which part.
Things learning AI can mean:
Chatting with an AI effectively to write software
Writing software that queries an AI for part of its operation
Running an agentic harness instead of chatting with the AI
Making use of models directly via hugging face (not API)
Training an existing model with existing weights
Doing machine learning starting with blank weights, and maybe unrelated to LLMs
Designing new AI architectures
All of that is learnable. But if someone says they want to learn AI, and someone says, ok, I'll help you, there is only a 1/7 chance they have the same slice in mind. But all of it is useful. And mastery of even one of those levels is powerful. No one needs to do all of those things.
It's kind of the first category. End use via chat. Maybe using to write code is a different level of power user. So I guess, it's somewhere around 7, 7.5, 8.
Linux navigation, beyond basic use (system config, coding, etc.), troubleshooting, self-reliance, confidence.
Wiki user/group management, addons, themes, etc.
Setting up out-of-the-box platforms on VPS, plus analysis. (ie. To set up a PeerTube or Invidious instance I'd want to keep an eye on bandwidth and such so I don't get a huge bill.) Then customization.
Log in
Coding AI programs. I am too old; it wasn't taught when I was in college, but it is now. 21-year-old graduates have an advantage over me.
The problem with AI and teaching it (very minor problem), is there are now so many layers to it. All of them are pretty teachable. But when someone says, I want to learn AI, there is an ambiguity in which part.
Things learning AI can mean:
Chatting with an AI effectively to write software
Writing software that queries an AI for part of its operation
Running an agentic harness instead of chatting with the AI
Making use of models directly via hugging face (not API)
Training an existing model with existing weights
Doing machine learning starting with blank weights, and maybe unrelated to LLMs
Designing new AI architectures
All of that is learnable. But if someone says they want to learn AI, and someone says, ok, I'll help you, there is only a 1/7 chance they have the same slice in mind. But all of it is useful. And mastery of even one of those levels is powerful. No one needs to do all of those things.
Also: business admin, writing/editing, graphics, filmmaking, and/or music generation. Or does that fit in one of those 7 categories?
It's kind of the first category. End use via chat. Maybe using to write code is a different level of power user. So I guess, it's somewhere around 7, 7.5, 8.
The ability to walk away and experience real life :P
Don't work for the man, quit and start up your own business, and work for yourself.
Amen. And I shall work in the sun and stop sitting.
Fucking A
https://youtu.be/wczkA_cULYk?si=6De3BD9xFlTO7Mt9POOP
Linux navigation, beyond basic use (system config, coding, etc.), troubleshooting, self-reliance, confidence.
Wiki user/group management, addons, themes, etc.
Setting up out-of-the-box platforms on VPS, plus analysis. (ie. To set up a PeerTube or Invidious instance I'd want to keep an eye on bandwidth and such so I don't get a huge bill.) Then customization.