If that was a genuine question, yes I have a very decent idea what prions are. We were discussing if the computational design tech should be giving people pause and the answer is no. If anyone really wants to create nasty stuff they dont need computers much and certainly no fancy ML tools. After the design, getting nasty stuff into humans typically happens either via viruses, pathogens, foods, or illicit drugs. Illicit drugs and foods are monitored in many parts of the world, and in Europe it is hard to mess with the food supply at scale. But even if you did, there is really no need for these designer tools as there are enough nasty sequences, including prions, that are known. On the other hand, these computational tools can help create safe biomolecules — it is just too easy to create nasty stuff with chemistry and biochemistry, but it is still very very hard to create safe and effective therapeutic molecules.

▲

lazide 19 hours ago | parent [-]

Viruses are unnecessary for prions (near as we can tell), and a nasty virus can do plenty of direct damage.

Someone could (if they discovered one) construct a nasty prion directly eh? And we’d probably be pretty fucked.

So far though, we only know of a handful of actual working prions, so maybe there isn’t other others?

	▲	pama 4 hours ago \| parent [-]
		> Viruses are unnecessary for prions (near as we can tell), and a nasty virus can do plenty of direct damage. Agreed > Someone could (if they discovered one) construct a nasty prion directly eh? And we’d probably be pretty fucked. It would be easy to construct a potent prion-like sequence that works in isolation (in vitro) with such a tool, but not clear what the rationale would be even if one has nefarious purposes. We know of enough of them (eg pick Creutzfeldt–Jakob disease for human hosts; if you have a hydrogen bomb, the variations on the engineering could be intersting but most likely not needed.) I dont think we would be fucked if someone constructed new prion-like sequences with this method, though we might be at risk if someone makes genuine efforts to find ways to deliver the existing (or new ones) to humans at scale. The main risk factors in prions diseases involve exposure and host compatibility. You need enough time and thus enough of a starting dose for harm to happen. You need to enter the host and go to the right place, say nervous system, without being chopped up along the way by random proteases and without self aggregating in the wrong place and becoming useless. There may be other computational tools that in combination would help realise more potent/nasty/risky designs, but it is probably much easier to use human cell lines and benchwork (design a funky evolution experiment) to fish out nasty sequences that actually work in human cells. Overall in biology and chemistry killing is easy, therapy is hard. Prions are somewhat interesting from a biology or engineering perspective, but they dont have the huge added risks of rapid exponential growth and easy transmission to hosts that many viruses have. They seem more like a curiosity and primitive danger form, and not a huge public health risk factor. Maybe there even exist some semi benign (even useful?) forms related to prions that we dont study much yet and they may end up influencing things like our understanding of aging or immune responses. Prions relate to nucleation and first order phase transitions, so they have some technical appeal to people with past background in physics or chemistry, including myself, but they are not making it in my top 20 threats for humanity, not even if there were concerted efforts by rogue actors to create new nasty ones.