pflanzenregal
Ich mag Pflanzen und hab ein extra Regal dafür. Support your FOSS (Free and OpenSource Software) projects!🫶🏼
- 0 Posts
- 12 Comments
1·1 year agoNailed it. No one can argue with that 🧠
Essentially because after they did something bad it’s likely too late. But others can explain this in more detail with more knowledge than I have.
This is a great example to explain it to people who are familiar with the topic. But if I tell that to a “random” friend, word for word, they won’t know what I’m talking about :D
Someone should make a post about why blocking Threads is good and why it’s not to be confused with gate keeping. If not properly communicated, this could look very badly for the uninitiated and they’re not to blame.
Some people of course have an educated opinion against blocking, but many presumably don’t know the reasons behind it.
3·1 year agoMaybe read abeut the EEE strategy (embrace, extend, extinguish). Gatekeeping is bad but this isn’t about gatekeeping
After a few times I memorized where the bread or fruit (w/o barcode) I usually buy is in the menu and am almost equally fast as an employee would be. So it just took me some time to adjust personally.
You don’t need a LLM for converting pseudo code to Latex. LLMs surely help at programming (in my experience), but I feel like your example is really giving them justice :p
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
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.
It’s better. Not good. Better than other tools, at least in the eyes of the many people using it. But as I stated at another post, to me this speaks to the fact that we need better FOSS alternatives for whatever purposes discord is used. I don’t like Discord either, don’t get me wrong! But so many people using it means something’s missing and I don’t think it cab solely be explained by the lack of knowledge of existing solutions but at least partly by the existence itself.