As a heuristic, I look at the commit dates to see if it’s still being maintained. And lately I look at the commits themselves for evidence of AI use. Then I’ll maybe do a web search for discussions to see what other people/the community has to say about it.

Gnuhealth looks like it may be dead (or at least the repo I found hasn’t had a commit in many years). (Nvm, they moved to codeberg which is a positive signal, IMO).

I think it’s illegal to use Reticulum over amatuer radio because it’s encrypted?

LoRa is very low bandwidth, only suitable for short text messages. Radio frequencies are heavily regulated, easy to triangulate, and can be shutdown. People could build a higher speed, planned WiFi mesh; hops just have to be closer. Neighborhoods could build cat6 and fiber into meshes if they wanted to as well.

I think he was downplaying his use. He said he was getting addicted to the “dopamine” hit it gave him, which sounds like he was using it more than he claims. I don’t believe his scripts weren’t at least partly written by AI.

Related Youtube video I saw the other day: “AI bubble will pop…………any day now.”

Companies boasting about committing felonies and trying to charge companies to run the uncontrollable super-hacking machines on their systems has got to be one of the stupidest marketing attempts ever.

“Can’t turn your brain off,” and not reading the docs seems contradictory to me. I.e. if you don’t even know the contract of the specific methods the code is calling, you don’t really know if it’s correct or not.

import anthropic

client = anthropic.Anthropic()

message = client.messages.create(
    model='claude-fable-5',
    max_tokens=128000,
    messages=[
        {'role': 'user', 'content': 'say "Hello World"'}
    ],
)

print(message.content[0].text)

I avoided all the effort of going to read the docs and relearn something I very rarely need.

This is the same thing as blindly copy-pasting StackOverflow snippets; which people used to think was a bad thing.

Half the living standards. Will be uncomfortable, and likely violent.

It’s meant to be an attempt at humor/sarcasm/engagement bait I think.

There’s a lot of loopholes. I think I read that it doesn’t affect Blackrock at all. It will affect Blackstone, but there are enough loopholes for them too.

I’ve done this (asked a frontier model red team my server), and it just gave me false alarms that I went investigating and weren’t really issues (would’ve been an issue if I had a very specific configuration, but not generally). They produce a lot of noise, so you still have to somewhat understand what you’re looking at/for. It didn’t find any legitimate issues, which is good I guess.

Governments are captured by these companies. Any regulation would undoubtedly benefit these companies at the peoples expense. Right now, the best solution is probably open research and open weight models to keep the field level.

In my experience (of having my office 2-doors down from the CEO of a minor public company), most of his “work” was just meetings with other CEOs and execs over breakfast, lunch, dinner, golf, etc to discuss business deals, acquisitions, and such. Not sure how an AI could do that, as it was mostly a social/class/nepotism thing.

Same (but only 6 months). Min wage + commission. But, during election season, because the company didn’t get commission, they just had the computers run through the lists and have us immediately hang up instead of going through the campaign scripts.

My favorite way is Borscht. They’re pretty good grilled, then cut up, and put in salad too.

Amazon was junk level for a long time. I don’t think it’s uncommon for early, “grow fast at all costs” companies to have poor credit ratings. Though, Oracle really isn’t an early stage company.

AFAIK, these companies are making record revenue.

The “best” one I’ve tried is the latest Opus. I don’t trust any of them to use for real work, so I mostly just play around with a local Qwen 3.6 27B or Deepseek v4 Flash. I have heard OpenAI’s latest models produce less bloat than Anthropic’s.

I should say I do find LLMs useful as a kind of search agent (both web and large unfamiliar code bases). And GhidraMCP is pretty cool (maybe just because I don’t have much experience with reverse engineering).