thanks

I’m really curious how someone can exploit a script that is meant to be running locally with no external facing interface

A mining rig should work, but you might want to restrict the GPU to one model, as loading 4G on PCIx1 when switching models would be too slow. Only one way to try it out really. At worst it could make an alchemist.

Separately do you know if AMD Radeon RX580s are capable of contributing to AI Horde?

Not currently I think

I also developed and am running the AI Horde but that’s not very fediverse adjacent (yet) :D

unfortunately since this goes via the pict-rs, there’s no such information. Hopefully the tooling will increase in the future.

However this tool is fuzzy by necessity. Most, if not all your hits will be false positive (because CSAM is actually rare). So you will need human review for that sort of approach

It’s the primary anti-CSAM protection in the AI Horde and it’s been running for months. I’ve done enough tests to be convinced it works, but I’m no scientist of course.

I’m hoping someone will do rigorous research on my approach at some point.

You can go one step further and use a conda env, which would also include the proper python version. All you need then is the micromamba binary. I might develop that as all it would need is to run a shell script to start

theoretically this tool could be adjusted to go via scp and read your filesystem pict-rs storage as well, Just someone has to code it.

As funny as this is, please don’t use this. It took a long time to get people to be vigilant about shady links, and this being in use anywhere will cause them to start unlearning it.

You couldn’t replace oracle with sun early on at least?

Those who had defederated via fediseer automated blacklist in advance should be clean.

I don’t think this makes much sense. You can’t synchronize upvotes/downvotes and posts. These belong to a different account.

What do you mean? It’s already setup.

About the gui, I’m not a frontend engineer. Hopefully someone can make onr

Someone vouches for a bad actor, bad actor vouches for more bad actors- then they can circle jerk their own reputation up.

There’s a chain of trust. If a bad actors lets in all their friend, withdrawing the guarantee from that bad actors, withdraws it from all their friends.

Also, what prevents actors in “downvoting” instances hosting content they just don’t like?

There’s no “downvote”, but even if I add it, I would add filter so you can ignore “downvotes” from people you don’t agree with, or only see “downvotes” from instances you agree with.

Actually, not a handful. Everyone can vouch for others, so long as someone else has vouched for them

For contributing, it’s open source so if you have ideas for further automation I’m all ears.

First we need to populate it. Once we have a few good people who are guaranteeing for new instances regularly, we can extend it to most known good servers and create a “request for guarantee” pipeline. The instance admins can then leverage it by either using it as a straight whitelist, or more lightly by monitoring traffic coming from non-guaranteed instances more closely.

The fediseer just provides a list of guaranteed servers. It’s open ended after that so I’m sure we can find a proper use for this that doesn’t disrupt federation too much.

I agree. This is why I started the Fediseer which makes it easy for any instance to be marked as safe through human review. If people cooperate on this, we can add all good instances, no matter how small, while spammers won’t be able to easily spin up new instances and just spam.

Absolutely. Me and a couple of others have been warning against this for a week now.

I noticellot of instances which were flooded with bots due to the open registration. I have most of them degenerated for this reason.