• 4 Posts
  • 254 Comments
Joined 1 year ago
cake
Cake day: October 23rd, 2023

help-circle










  • One suggestion is to look at getting any useful docker container up and running first to get a feel for it, then pivot towards heavier lifts. Self-hosting emails is going to be a headache as others have said, and monero nodes sounds like it could be a technical as well but that’s admittedly an uneducated guess.

    You can look at the recent selfh.st poll for the most popular docker containers, see if any strike your interest and start there.


  • There is an inherent catch-22 in that the people who believe AI a fluke will rarely have the time/interest in realizing the vast improvements since the initial launch 2-3 years ago.

    I am as frugal as they come, yet I shell out money for the paid version and have a reference point on its helpfulness that is grounded in experience. It is almost impossible trying to create a reference point to those that have no (recent) experience and refuse to get any.

    It’s human nature/Dunning-Kreuger so I can’t be too frustrated I suppose.


  • Troublehsooting technology.

    I’ve been using Linux as my daily driver for a year and a half and my learning is going a lot quicker thanks to AI. It’s so much easier to ask a question and get an answer instead of searching through stack overflow for 30 minutes.

    That isn’t to say that the LLM never gives terrible advice. In fact, two weeks ago, I was digging through my logs for a potential intruder (false alarm) and the LLM gave me instructions that ended up deleting journal logs completely.

    The good far outweighs the bad for sure tho.

    The Linux community specifically has an anti-AI tilt that is embarrassing at times. LLMs are amazing, and much like random strangers on the internet, you don’t blindly trust/follow everything they say, and you’ll be just fine.

    The best way I think of AI is that it’s going through a bubble not unlike the early days of the internet. There was a lot of overvalued companies and scams, but it still ushered in a new era.

    Another analogy that comes to mind is how people didn’t trust wikipedia 20 years ago because anyone could edit it, and now it is one of the most trusted sources for information out there. AI will never be as ‘dumb’ as it is today, which is ironic because a lot of the perspective I see on AI was formed around free models from 2023.