brucethemoose

joined 6 months ago
[–] [email protected] 6 points 4 days ago* (last edited 4 days ago) (1 children)

The behavior is configurable just like it is on linux, UAC can be set to require a password every time.

But I think its not set this way by default because many users don't remember their passwords, lol. You think I'm kidding, you should meet my family...

Also, scripts can do plenty without elevation, on linux or Windows.

[–] [email protected] 27 points 6 days ago

That actually is weird.

[–] [email protected] 1 points 1 week ago* (last edited 1 week ago) (1 children)

The problem is that splitting models up over a network, even over LAN, is not super efficient. The entire weights need to be run through for every half word.

And the other problem is that petals just can't keep up with the crazy dev pace of the LLM community. Honestly they should dump it and fork or contribute to llama.cpp or exllama, as TBH no one wants to split up LLAMA 2 (or even llama 3) 70B, and be a generation or two behind for a base instruct model instead of a finetune.

Even the horde has very few hosts relative to users, even though hosting a small model on a 6GB GPU would get you lots of karma.

The diffusion community is very different, as the output is one image and even the largest open models are much smaller. Lora usage is also standardized there, while it is not on LLM land.

[–] [email protected] 1 points 1 week ago

TBH this is a great space for modding and local LLM/LLM "hordes"

[–] [email protected] 7 points 1 week ago

^

Futurama had it right, spammers are the ultimate destroyers.

[–] [email protected] 7 points 1 week ago

Then the lemmy title is misleading, no? Isn't that against the rules?

[–] [email protected] 3 points 1 week ago

Please ask him, tape it, and don't let the campaign managers talk him out of it.

[–] [email protected] 1 points 1 week ago

+1

Never attribute to malice that which is adequately explained by wanting to make money.

[–] [email protected] 1 points 1 week ago

You wouldn't steal a car...

[–] [email protected] 1 points 1 week ago

Hmm, what if the shadowbanning is 'soft'? Like if bot comments are locked at a low negative number and hidden by default, that would take away most exposure but let them keep rambling away.

[–] [email protected] 29 points 1 week ago

Top 50% of the population still.

After all, they wrote a review.

[–] [email protected] 34 points 1 week ago (2 children)

Trap them?

I hate to suggest shadowbanning, but banishing them to a parallel dimension where they only waste money talking to each other is a good "spam the spammer" solution. Bonus points if another bot tries to engage with them, lol.

Do these bots check themselves for shadowbanning? I wonder if there's a way around that...

view more: ‹ prev next ›