Skip to content

Self Reproducing Machines at the Berlin Desktop Summit

Wednesday, 24 August 2011  |  richard dale

The Berlin Desktop Summit was great, and I think it is about time I wrote something about my thoughts.

I look forward to software conferences because you can never tell which ideas will excite you most. In 2011 I would expect to be wowed by the latest in tablets or 3D rendering stuff, but actually it turned out that 3D printers and a bad attempt to build a toaster from first principles were what left me with the deepest impression at the Summit. Michael Meeks gave a Lightning Talk on his 3D printer RepRap project. It was really funny, about how he built five iterations of his printer, with each generation printing the next printer. Sadly it seemed his wife's nylons suffered in the cause of science, but holy crap I'd personally happily donate my socks to further such an awsome project. I don't know how Michael does it, I am a big fan of his blogs where he describes his thoughts on software such as Libre Office, massive child rearing efforts, attempts to fix his plumbling, lots of stuff on learning Christianity and of course those 3D printers. I couldn't actually write a blog like that because I personally manage to do bugger all apart from mainly writing software, listening to music and drinking a lot of beer, and if I wrote about my life, by comparison sadly it would be a bit of a dull read. Oh well. I can only think about one thing for years on end it seems, and I wish I was more of a generalist like Michael. But if I wanted to think about one thing, there couldn't be many better topics than self replicating machines.

Thomas Thwaites gave a talk on how he built a toaster from first principles was equally great - he was entertaining while at the same time making you think. He described how you could buy a toaster for about 10 euros, which incoporated all sorts of advanced technologies that a single individual or company could never manage to discover by itself. A 10 euro toaster depends on an eco-system that no one company or individual understands. I read a similar account of how a pencil couldn't be constructed without a large inter-dependent eco-system, but I can't find the reference. The point is that it is very, very hard to produce the simplest of the things we use without a lot of advanced technology and cooperation, that has evolved in human society over the past five to ten thousand years or so. Thomas eventually built a really bad toaster, and it doesn't matter that it was bad because the important thing was that he was able to uncover some of the assumptions we make about what it takes to build things from scratch.

The connection between Michael and Thomas's talks is in relation to thinking about how we can go about constructing self replicating machines. See this Self-replicating machine article in Wikipedia. If we want to send a machine to a distant planet, and have it create self replicating machines we need to be able to extract natural resources like mica and iron and steel like Thomas described, and then use them to build machines which can then reproduce themselves, extract more resources from the host planet to use for building better copies of themselves. The awsome thing is that I think with open collaboration via the internet, with open sciencific method principles not hobbled by patents, we can collectively achieve a self reproducing Von Machine within a relatively short timescale (under 50 years). Woo hoo!

CORRECTION: it wasn't Michael Meeks wife's stockings that he needed for extra raw material, it was a nylon kitchen chopping board I think.