Evolution cannot optimize for maximum harm. Natural viruses only evolved to reproduce, with the harm just a side effect. Humans can combine features of multiple natural viruses. There's no obvious reason why we couldn't make something with both the infectivity of measles and the lethality of rabies, as well as resistance to all known antivirals. If something like that gets out it's probably civilization ending.

Because it massively raises the amount of damage that a lone-wolf crazy person can do.

Not to mention causing a famine or lack of medicine by sabotaging the cold chain.

What if the white supremacists make a retro virus, that for instance only kills people with sickle cell anemia..

Are you actually asking this? Like why should people be worried that a disgruntled highly motivated person would venture into this area? Because it can happen. The outcomes of it could be really bad? It also doesn't have to be done by an individual, could be nation state or organization led. There are groups of people who believe humanity needs to end and all sorts of crazy stuff. Some of them have gone on mass murder suicide missions.

2 bad things can be threats at the same time. Even 10,000 things can be.

That seems a bit silly thinking they are not doing anything to protect the water supplies, and even more silly to not compare the impact of the things you are comparing. Poison a water supply and you hit how many, a city, a small part of a city? Now compare it to the impact of a virus like corona - we literally just went through that..

Coronavirus++ is not something that I would want on my calendar

Disrupting electric supply, water supply, and food supply chains are already tried and tested separately, with some limited monetary system disturbances, here and there. So doing those simultaneously - possibly leveraging on latent vulnerabilities kept covered for later abuse - sounds a natural step 'forward', close to trivial for those into that. Causing huge broad disturbances and possible unrest in societies. So that some governments started to ask people to build up reserves.

I can imagine that terrorist are sad now that they cannot use their preferred advanced bioengineering skills acquired through great efforts posing as scientist towards the so far unsuspecting services providing them with harmful DNA building blocks and now have to turn to much much more pedestrian methods instead of the heightened sophistication they crave.

Or to sources from other countries.

Its hard to imagine a mustache twirling terrorist bent on putting PFAS and lead in the american water supply, putting mad cow in the beef supply and making everything cost more because he would have been outcompeted by the US government.

Your status as a "defender" is measured by the fanciness of the threats you defend against, not their actual magnitude. That's why Space Force is so popular, and everyone working on vaccines is miserable.

Also, see the "lab leak" discourse. To me the question seemed irrelevant because the countermeasures would be exactly the same, but that's not how other people think; they're not interested in a mask mandate against a naturally occuring virus, but if they can tie the exact same genetically identical virus to a human enemy? The social status of the threat goes up!

Regular old measles? Not sexy. But a convenient vector for the internal threat. https://www.politico.com/news/2026/11/08/americas-measles-ep...

Made-up AI-assisted super-measles? Doesn't exist yet. But it's a great super sexy threat to defend against. You just have to ignore the kids dying of boring old regular naturally occuring measles in the meantime.

Lots of scary stuff has been possible for ages. Much of it is so easy that wouldn't be publishable because of lack of novelty.

I see LLMs as being essentially irrelevant for whether someone can do this. It's lab skills, planning skills, biological experiment debugging skills.

"Why did my plasmid not start getting expressed, how did I screw up?" "How do I make this huge protein" aren't LLM questions, they're university level biology questions, and "why is my protein growth tank not making the thing I want, and what is the grey stuff, do I need to put clean it again" is an experience question.

Putting checks on DNA synthesis puts making the horrible stuff out of the range of what high schoolers can do though.

Well look at what it's doing to the hacking scene. It doesn't have to do all the work for you but people knowledgeable and motivated enough are now 10-100x more of a nuisance.

It's not like you can prompt it to create super AIDS in an afternoon but I'm sure an uncensored model could help you do real harm.

It's more likely they're surfacing the contents of deeply technical biology papers to average people who may or may not be thinking about terrorism.

It doesn't seem to be? From what I can tell, this is just formalising what we've had to self-declare with companies like IDT for years (not encoding a known toxin, pathogen, etc). Somebody correct my if I'm wrong, but it looks like there's now a requirement for companies to actually check the sequence, rather than take the researcher's word for it.

You don't need AI to generate known toxin or pathogen sequences (as these are already known).

And if AI could design novel toxins, this screening wouldn't catch those anyway.

Alphafold or something like that is more likely than the LLM models. Though cant say what an LLM model that uses Alphafold as a tool can do.

No not at all. The top models are not uncovering anything novel with DNA.

This is old stuff, being implemented and getting lots of news.

Probably a good thing overall in my opinion, even though it obviously has impacts.

I got into this hobby briefly (pre-LLMs) before realising how much of a loaded gun it is, and that I’m just not qualified to safely muck around with this stuff.

I suppose you might have to go back to bathtub explosives as a safer alternative.

Yeah me too. Using Clorox on my toilet seat kills always 99.99% germs. The fun is with the 0.01% that's left

The device used to make oligonucleotides is pretty much an inkjet printer, from there it's cleverly applied enzymes to assemble the full strand.

Usually the synthesis of oligonuclotides is done with phosphoramidite chemistry [0]. So tell us more about those enzymes to make DNA molecules from scratch.

[0] https://www.twistbioscience.com/blog/science/simple-guide-ph...

>So, a song and dance of doing something that will most likely just inconvenience people doing real work.

Why does he call it an "otherwise pretty great policy"?

He's claiming that the policy needs to be more comprehensive. You're concluding that it doesn't pass a cost/benefit test. That's not at all the same claim!

My guess is that the assumption is that US gov funded labs are enough of many suppliers revenue that it'll be effective, but I'm 100% speculating and it may be total garbage speculation given that I have no idea how much hobbyist level DNA projects are a thing or how much they're a thing for non-medicinal business purposes.

If you read the next two sentences it makes it fairly obvious that this is a plan to indirectly target suppliers. Federally funded researchers will be required to

"procure synthetic nucleic acids and benchtop synthesis equipment from U.S.-approved suppliers. Those vendors must verify their customers’ identities and screen orders for potentially dangerous sequences that, for example, could encode toxins or pathogens."

If this was just about feds self-policing, they wouldn't have the customer identity requirement.

Probably they're taking this route because AI is moving fast and Congress moves slowly.

Trouble is, that assumption hasn't always worked out in the past: https://en.wikipedia.org/wiki/Bruce_Edwards_Ivins

But yes, this is stupid. No one who is inclined to build a biological WMD is going to be more than mildly inconvenienced by supply restrictions. DNA isn't plutonium.

When DNA is outlawed only outlaws will have DNA.

Yeah, I'm assuming that was intended to be sarcasm.

(Also, my comment is great, because I wrote something instead of nothing.)