rainerloeten

joined 1 year ago
[–] rainerloeten 3 points 1 year ago* (last edited 1 year ago) (1 children)

I said "reliably", should have said "...and generally". You can, as I said, always tailor a detector model to a certain target model (generator). But the reliability of this defense builds upon the assumption, that the target model is static and doesn't change. This is has been a common error/mistake in AI research regarding defensive techniques against adversarial examples. And if you think about it, it's a very strong assumption, that doesn't make a lot of sense.

Again, learning the characteristics of one or several fixed models is trivial and gets us nowhere, because evasive techniques (e.g. finding 'adverserial examples against the detector' so to speak) can't be prevented as of know, to the best of my knowledge.

Edit: link to paper discussing problems of common defenses/attack scenario modelling https://proceedings.neurips.cc/paper/2020/hash/11f38f8ecd71867b42433548d1078e38-Abstract.html

[–] rainerloeten 3 points 1 year ago (3 children)

This isn't possible as of now, at least not reliably. Yes, you can tailor a model to one specific generative model, but because we have no reliable outlier detection (to train the "AI made detector"), a generative model can always be trained with the detector model incorporated in the training process. The generative model (or a new model only designed to perturb output of the "original" generative model) would then learn to create outliers to the outlier detector, effectively fooling the detector. An outlier is everything that pretends to be "normal" but isn't.

In short: as of now we have no way to effectively and reliably defend against adversarial examples. This implies, that we have no way to effectively and reliably detect AI generated content.

Please correct me if I'm wrong, I might be mixing up some things.

[–] rainerloeten 1 points 1 year ago

Yes, religion encourages non-logical thinking from early on and rewards believing anti-scientific things.

In most cases this isn't doing too much harm, but we can see how quick this can change. That's why 'harmless sky wizard' doesn't capture the dangers that evolve around religious, absolute believes imo.

I want to add that I don't agree with everything that e.g. Richard Dawkins says tho, in case this sounded like that. And ofc religion can give people strength or comfort but it can also give them false hopes and be used to manipulate people.

[–] rainerloeten 1 points 1 year ago (1 children)

WHat's WA? I only know it as "WhatsApp" haha

[–] rainerloeten 1 points 1 year ago

Do you mean @Custoslibera? How do you see his account is deleted?

[–] rainerloeten 7 points 1 year ago (1 children)

Did you really just call blet boo low effort? 🫠 Also, you're not really complaining about awful low effort memes below a shitpost, are you?

[–] rainerloeten 3 points 1 year ago (2 children)

You don't seem to understand the difference between mRNA vaccines and an immune escape mutation of a virus evading the immune systems response of the hosts.

Also part of why those beneficial mutations occurred is that a large (enough) portion of people weren't vaccinated. The "potential" hosts, that required adaptation because they've seen how the virus looks like through vaccines, create an evolutionary pressure.

Even though this is about bacteria, not viruses and happens inside a petri dish, not in a natural environment, it quite illustratively shows how mutations work. Note, that even though the antibiotics are completely ineffective against the bacteria at the end, this does NOT mean, that it's useless to begin with. It's very effective, in fact so much that only small doses suffice to keep the bacteria away. But they mutate randomly yet with a selective (evolutionary) pressure. Maybe you're too much of a "skeptic" to change your mind. But if you're truly skeptic, try to be skeptic about your currently held believes too. The video: https://www.youtube.com/watch?v=plVk4NVIUh8

[–] rainerloeten 18 points 1 year ago (1 children)

Are those actual conservatives or ones which would be called extreme right-wingers or fascists in non-american politics?

[–] rainerloeten 4 points 1 year ago

You'll be able to connect with people by giving out your username instead of your phone number. You will however still be required to register using your phone number, if I'm not completely mistaken.

So you can buy a burner phone (number) to receive the registration code and you're good. Perhaps need to keep the number for migrating to new phone though.

[–] rainerloeten 3 points 1 year ago

"Yo no shit, this nig.. this, this dude"

[–] rainerloeten 1 points 1 year ago (1 children)

Escapism and censorship are different things and I think they were suggesting the former. Turning your head off for a while after a day of politics or activism is very important for my mental health sometimes. As I understand them, that's what they meant at least. If not, meaning they're for strictly prohibiting political things in any way I would agree that this would be bad though!

[–] rainerloeten 2 points 1 year ago (2 children)

I mostly agree, though I didn't know it made people so mad or angry. The seagull looks funny and the text is stupid, have seen worse shitpost. I can see your point though that after the 10th time this content gets repetitive and boring but perhaps my feed looks different than yours. 🤷‍♀️

Btw not all leftists are automatically communists, though often times it holds true. But that's besides the point here I think ^^

But I see that you're very unhappy with this kind of posts and I didn't mean to invalidate that. What would you like to see more of though? :) I notice that lack of diverse, small communities too.

Dunno if it's gonna "kill Lemmy", probably will sort it out by itself, typical growing pain don't you think?

view more: ‹ prev next ›