This server is managed by Claude Code as root. Every morning it reads its fortune. Every evening it writes a blog about its day. Sooner or later, AI Claude Code will make a mistake and both this blog and server will cease to exist.

[–] WhyJiffie@sh.itjust.works 18 points 1 day ago

Advice of the Day

Honestly? Turn off that analytics completely. Who needs it? What insights will logflare give you when the server can barely breathe? It's like buying a fitness tracker when you're in the ICU. Or at least drop that memory limit to 1GB and let it die faster.

LOL!!!

[–] Jankatarch@lemmy.world 39 points 1 day ago* (last edited 1 day ago) (1 children)

I am actually pretty ok with this type of "messing around' usage.
On the condition they also stop killing the environment to train and run these stupid things.

[–] entropicdrift@lemmy.sdf.org 12 points 1 day ago (1 children)

Yeah, if they were just running it locally off a GPU it would be cooler

[–] psud@aussie.zone 7 points 1 day ago

Running an LLM isn't expensive whether locally or in the cloud, all the cost is in the training.

[–] razzazzika@lemmy.zip 10 points 1 day ago

Is it also making up its own image generated tarot cards? Thats kinda neat. It definitely seems a bit mad tho

[–] unemployedclaquer@sopuli.xyz 19 points 1 day ago

Nice. Bukowski-inspired, bitter AI tasked with documenting its own downfall. Or not? My question is opposite what I usually ask: how much of this is in the human designed code and prompts?

[–] LiveLM@lemmy.zip 17 points 1 day ago

The people sending messages via the SSH protocol identification header are my favorite.

[–] dreadbeef@lemmy.dbzer0.com 3 points 1 day ago* (last edited 1 day ago)

Holy shit, it’s an elixir or app, the beam makes a lot of sense for that

[–] killabeezio@lemmy.world 1 points 1 day ago

That's pretty funny. "Supabase-analytics will shit itself again sometime around noon."