Coz we a lazy bunch. Deal with it.

The AI is just image and object recognition and tagging. It’s very powerful (even runs on CPU in docker) and useful. No LLMs here.

Yeah, I didn’t get why they would think it’s not stable.

Context: https://lemmy.world/post/39283431

Why? Don’t they use cardboard straws for their frappes?

I mean, maybe you like rawdogging life. I dunno.

My network shared folders are on a windows 11 (yes, I know. It’s shit.) pc and my docker is running on Linux.

Here’s what my mounts look like -

volumes:
  plex:
    driver: local
    driver_opts:
      type: cifs
      o: username=pc_username,password=pc_password,vers=3.0,file_mode=0777,dir_mode=0777
      device: //10.0.0.3/Plex

Hope this helps.

Folks, tell me if this is a good idea - OP gets a backblaze subscription. Backs up everything on that system - all the forgejo stuff, all the immich stuff, all the Arr content.

If/when stuff breaks, OP… gets a backblaze drive home with their stuff and returns it after reinstating their backups?

Truth or dare?

Truth this time.

What’s your credit card info?

Didn’t read the article, as you said it’s just an ad for VPNs. But one way this affects VPN users is that now you can’t VPN into Italy to use a service without hitting the age verification prompt. That just sucks!

Sadly, true.

Hear me out. After they get fired, other essential workers should also go on strike. Every. One.

I wish the air traffic controllers would go on strike. Two days of that and suddenly congress would open right back up.

Yeah, they really need to start building RAG supported models. That way they can actually show where they’re getting their data, and even pay the sources fairly. Imagine a RAG or MCP server connecting to Wikipedia, one to encyclopedia.com, and one to stack overflow.

It’s not easy. LLMs take so much training data that at this point, their training data is basically, all books publically available, all blogs on the internet, pretty much all of tumblr, Reddit, stack overflow and every forum you can think of. Even then, some LLMs need even more data. So companies have started outright stealing data - pirating stuff, downloading stuff from Anna’s Archive, etc.

So no, no billion dollar company can make their own training data. Even if you plug in every email ever sent on Gmail, Google still won’t have enough data to train a good LLM. So they go with the cheaper option- training data that has already been collected, sorted, cleaned, and labeled.

In one sense, they’re again stealing others’ hard work - rather than cleaning their own data, they use public data sets. In another sense, even that’s not enough.

OP, you mentioned Mouse as an example in a comment. You’re on the right path - it would have been trivial for Cypher to program (or have Mouse program for him) a mini Matrix where he’s uber rich and everything he wanted. Of course, it would get old real soon.

I reckon Cypher would even have mentioned it a few times to people… things like “man, I miss meat” or “man I wanna be rich af”. Most would have taken his words as just bluster. Others would equate it with Mouse’s actions - a coping mechanism or just a temporary relief.

But the reality of it - only the machines control the real Matrix. Only they decide what gets in and out, barring a few hacks like the one that got Neo out and the others out.

To plug a full fledged human back in would require quite a lot of work. As the mania of wanting back in would have gripped him, Cypher would have left clues back in the Matrix and the machines, as someone else said, would have groomed him to betray.

Trump’s return to the White House this year has brought a revival of his adversarial relationship with the media.

Keep calling a nuclear bomb a stick of dynamite. That’s how the news will die.

Bruh. JSON.

My SO was told to travel to office every day of the week, only to sit in zoom meetings because all of their team is elsewhere.

Reaaaal good use of everyone’s time and our non-renewable resources.