Day 3/30 Underrated Ideas in Biotechnology In 1948, John von Neumann gave a series of lectures about “self-replicating automata.” His idea was to send self-replicating machines out into the Cosmos to explore distant planets and spread through the galaxy. Von Neumann’s automata had a few different components: a “universal constructor” would, given a tape encoding instructions to build a copy of itself, grab parts from a stockroom and “assemble them into a copy of itself.” A separate module, called the “universal copier,” would duplicate these instructions and pass it to the next machine, thus activating it. Obviously, this is unlikely to happen with metal and bolts. There are no man-made machines able to do this. But biology can! Cells carry genomes which they pass onto offspring. Cells also assemble themselves by harvesting atoms from their nearby surroundings. Cells are thus self-replicating automata and, furthermore, we can accelerate them to distant planets. In 2022, in what I think is one of his most underrated papers, George Church wrote a single-author paper exploring this idea. "Living cells on Earth…perform functions, such as replication from only simple chemical inputs," he writes, that are "impossible for all current human-made machines." Church imagined that bioengineered cells (perhaps dormant spores, which can persist for thousands of years and then “reawaken” later) could be carried on tiny lightsails. Most probes will be destroyed from comets, dust, or debris, so we’d have to launch trillions of them to ensure plenty reach distant planets. He calculates that, for a journey of 4×10¹⁶ meters, astroprobes would collide with about two dust grains on average. Any hit from dust would likely be fatal, so the survival of any given astroprobe is between 10-20%. Still, Church does cost estimates and finds that for *less* than the cost of a 1,000 gram-scale Starshot launch, you could instead launch 10¹⁵ picogram-scale probes. Probes that survive space debris and dust would crash land on distant planets, and then use carbon and other local atoms to rebuild themselves and divide. These cells could be engineered to biofabricate more astroprobes and lightsails. Perhaps they could even be engineered to make a “communicator” module that sends messages back to Earth. This sounds insane, of course, but here again Church has an idea: “The communications ‘device’ could be constructed and aimed using engineered organisms through…planetary-scale bioluminescence,” he writes. He imagines these cells could divide and colonize vast swaths of the planet, and then coordinate bioluminescent flashes (perhaps using synthetic gene circuits) that are “brighter than the resolution-limited surroundings” and thus can be seen using the James Webb telescope. By accelerating one of these picogram-scale probes to 5% of light speed (which is totally doable) they would reach Alpha Centauri in ~100 years. At 15% light speed, they’d get there in ~30 years. This paper goes into a lot more detail about how big each lightsail might be, how to launch them from balloons, and so on. I’d love to see some early, experimental steps toward this vision.