Recent submissions

Comment · 8 hours ago
Maybe it'll force the internet to go back to analog.
Comment · 4 days ago
I'm not talking about outcome, but of sensing. I also never wrote it was a suitable method. When did that become a requirement?

I kist thought it was interesting that there are receptors that can detect wetness, just not on the skin. It was a joke pointing that out.

Fuck.

Comment · 4 days ago
Secondhand smoke does not feel the same as drowning.
Comment · 4 days ago
Im a vacuum you would not be choking.
Comment · 4 days ago
Trying to beathing in a vacuum would not feel the same as breathing in a liquid.
Comment · 4 days ago
Your lungs have sensors that can detect being coated by a liquid which invokes coughing. In that sense, you can detect wetness.
Comment · 4 days ago
Oh, so if you stick your head in a liquid and breathe you wouldn't know?
Comment · 4 days ago
Stick your head in it and breathe. You'll know.
Comment · 8 days ago
I see what your saying. Yes, I suppose it's possible, but you couldn't test it to see if it worked until both entire models were trained and released. You'd also need a lot of poisoned data to get it to work, there'd be evidence of the poisoned data left for everyone to see, and you can't screw up the models performance too much, since it would be benchmarked.

Simply downloading a model from huggingface, adding a feature, such as abliteration, and uploading back doesn't face such hurdles.

Comment · 8 days ago
But anyone can decompile it, now, too.
Comment · 10 days ago
I don't think so. Both models would have to get into a blackhat's hands, who would then poison them both so if a keyword was present, the model giving the response would insert an exploit, and the sentry model that's supposed to check would ignore it.

So if you didn't have a checking model, you only need to download one model that the blackhat corrupted to fall victim. If you have a checking model, you need to download two models. So it goes from x/n to a x/(n*2) chance of being compromised. Furthermore if you downloaded a clean checking model, you would be alerted if the model giving the responses did something cheeky. Then you could warn others, so that makes it a lot harder for a blackhat to create poisoned models that would go unnoticed.

Comment · 11 days ago
Well you could use a model created from a completely unrelated entity. It would be difficult for an adversary to poison both in a exactly complementary way.
Comment · 11 days ago
Not to mention, Big Bathtub.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton