DNA-based chemical compiler

arxiv.org

63 points·lainon··8 comments

Ok, that is amazing. How long before llvm has dna support?

Once that happens you will be able to make a self hosting dna builder, and have a bunch of high level languages along for the ride :)

Hmm. On the one hand that’ll be great for cosmetic surgery and furries (grow compatible but different body parts in vitro); on the other hand, do we really want to live in a world of constant security patches for our DNA because someone’s found a new vulnerability that hackers can exploit? Or worse, CSS for limbs?

Or worse, have to deal with viruses...

Oh, wait....

Natural viruses aren’t malicious; they even have evolutionary pressure to become self limiting. Artificial ones, not so much.

Natural ones also appear far more infrequently than computer viruses, and I’d expect the same pattern with artificial organisms.

The same is true for computer viruses! Being a huge menace to your computer host is a good way to get yourself deleted. The best computer viruses are careful and spread silently.

And in both cases I'm sure we haven't even found the most silent virus there is.

Isn't the title a bit misleading?

> Hence, our attempt to convert high-level code to assembly code takes us one step closer to completing the dream of making a chemical compiler.

This reminds me of Nuskell, a chemical reaction network to "DNA domain-level strand displacement" compiler build by Stefan Badelt et al in Erik Winfree's lab at Caltech.

https://github.com/DNA-and-Natural-Algorithms-Group/nuskell

Reply on news.ycombinator.com