Maybe that’s the problem. Maybe it’s a guy who thinks the replicators on Star Trek are real. Maybe it’s another article from the delusional weirdos of the Singularity Hub. Maybe it’s just that I get really annoyed with physicists who think they understand biology. But yeah, Thomas Hornigold believes that we’ll be able to make desktop replicators that will make anything you want.
These tiny factories will be large at first, like early computers, but soon enough you’ll be able to buy one that can fit on a desk. You’ll pour in some raw materials—perhaps water, air, dirt, and a few powders of rare elements if required—and the nanofabricator will go to work. Powered by flexible photovoltaic panels that coat your house, it will tear apart the molecules of the raw materials, manipulating them on the atomic level to create…anything you like. Food. A new laptop. A copy of Kate Bush’s debut album, The Kick Inside. Anything, providing you can give it both the raw materials and the blueprint for creation.
Just copy biology! It’s not physics, it’s got to be easy!
In recent years, progress has been made towards this goal. It may well be that we make faster progress by mimicking the processes of biology, where individual cells, optimized by billions of years of evolution, routinely manipulate chemicals and molecules to keep us alive.
All we need is energy from the solar panels we’ll build with our replicators to power our replicators!
Suddenly only three commodities have any value: the raw materials for the nanofabricator (many of which, depending on what you want to make, will be plentiful just from the world around you); the nanofabricators themselves (unless, of course, they can self-replicate, in which case they become just a simple ‘conversion’ away from raw materials); and, finally, the blueprints for the things you want to make.
Let me just point out some basic biological realities.
Biological machines are not generic synthesize-anything machines. Enzymatic reactions are narrowly specific: they require very specific inputs (not just a bucket of dirt) and they are honed by evolution to produce very specific output — not just a particular molecule, but a particular chiral form of that molecule. There are very few general, ‘programmable’ molecular machines — ribosomes come to mind — but that’s only going to be useful if you want to produce proteins. Proteins are remarkably flexible, but still, they’re not sufficient if you want to make solar panels, or batteries, or a car.
He trivializes the difficulty of making the ‘blueprints’. I presume he’s thinking of genes, which are not blueprints, and that you’ll just be able to feed in the ‘language’ of your replicator, and it’ll build something complex for you. We don’t understand all the processes that build a cell, so that’s a long way off, and you have to consider the nature of what organic processes assemble. Are you going to build a cell phone made of meat, or wood, or chitin?
He’s also trivializing the energy requirements. You’re going to have to provide your bio-replicator with chemical energy — or you’re going to have to include a fairly complex mechanism for transducing electrical or light energy into chemistry. It’s doable, cells do it all the time, but it’s still a rather elaborate process with more energy losses.
And then there are the raw materials. Water & dirt? You mean organic carbon, phosphorus, nitrogen, and oxygen and hydrogen — fertilizer and gases and water. Can you grow a stalk of wheat in your cubicle? If you can’t do that, what are you doing babbling about the far greater task of fitting a whole farm, fields and livestock, plus an electronics factory, plus an IT department, all into a box on your desk, with negligible requirements for energy or feedstocks. And it has to come preprogrammed with the capability of synthesizing anything.
Singularitarians. They’re the 21st century version of happy clappy religious fanatics.