gfish: (Default)
gfish ([personal profile] gfish) wrote2011-06-13 10:10 am

3d printing

Being a big fan of near-future singularity fiction, I've been following 3D printers for some time now. Particularly the open source designs like the RepRap. Anything that brings something designed in a computer into the real life has an inherently fascinating touch of the occult to me. One that can brings parts of itself into the real world, von Neumann style, is even better.

The problem is, as intellectually neat as they are, the final product they're capable of printing just isn't very good. I'd love to have one, and I've thought about it, but I'd really just be getting one for the experience, not for practical use. I can always go rent time on one at Metrix, if I do ever have a problem that can only be solved by an ugly little piece of plastic. Given my comfort working with nicer materials, that's just not a common occurrence -- it has yet to happen, in fact. Hot metal smells way better than hot plastic, after all. Without a real use for one, I might crank out a couple sets of brackets, just to claim self-replication status, but then it would end up sitting unused. I'd rather just follow the blogs and watch other, more motivated people work out the problems and slowly improve the system. Neat idea, but just not very practical.

That was my mental status on the subject earlier this year, when I re-visited the Computer History Museum in Mountain View and finally got to see the full set of exhibits there. Not being a stranger to the subject, I was well versed in the early history of home computing. But seeing the artifacts in person, all laid out together, made something click in my mind. Home computing in the 70s was an absolute joke. A real company once had a "kitchen computer" for sale for $20k -- and the only output it had was a set of binary lights! (This was the earliest recorded incidence of "you can keep your recipes on it" marketing.) The hoemmade units weren't any better, of course, just much cheaper while being even harder to build and use. It was all very neat stuff, of course, and I very likely would have been keenly interested at the time. But completely useless and very easy to dismiss. Sounds familiar?

Not a shocking revelation, I suppose. But it was odd to think about a modern set of technologies as possibly being in the same situation. It's not hard to invent scenarios in which home 3D printing is amazing, but it is certainly hard to see how a RepRap or MakerBot or even a high end commercial unit can get there. And maybe the analogy is false. Physical objects tend not to scale as omg well with time as electronics do. The inherent physical limits are already a lot closer. I'm still not going to get a 3d printer any time soon, for all the reasons listed above. I've got more than enough interesting hobbies at the moment, thank you very much. But I'm definitely left wanting to keep an even closer eye on what is happening there.
anansi133: (Default)

[personal profile] anansi133 2011-06-13 07:38 pm (UTC)(link)
I suppose my latest fascination, papercrafting, is even less durable than the plastic that's less durable than metal. A papercrafted craft robo is theoretically possible, but it wouldn't last very many jobs.

So too for a Lego machine that can recreate itself from Lego bricks. At some point you gotta go out and buy the feedstock for these things.

The metal lathe is the first such tool I ever heard of that was capable of self reproduction. it doesn't sound so impressive if you factor in the human attention necessary for birth.
solarbird: (Lecturing)

two contextual points

[personal profile] solarbird 2011-06-13 07:34 pm (UTC)(link)
There are two bits of context that I think the 3D printer (and similar) hackdoms are lacking, and that's kind of a shame, because both contextual components provide really, really strong impetus towards crazy amounts of work with little obvious reward.

First, computers were magic in the way that parts fabrication - even for awesome projects - aren't. Look at computers in SF, computers in popular culture, computers in the popular imagination before 1980. At the more realistic end, you see things like HAL - or, very rarely, extreme moments of mostly-sanity like in Space: 1999, where its computer was a lot more like, you know, a computer. But mostly it's Colossus or the Enterprise computer, where computers were giant brains with magical knowledge that could answer any question and maybe even come alive. Benign computers like on the Enterprise could do amazing things, you just needed to know how to tell it. Evil computers took over the world. Because MAGIC. And even though the hackers knew better, it was still informing attitudes.

Then the two most important computers in history came out: the IBM PC (turned computers into stupid, boring office machines) and the Commodore 64 (turned computers into stupid entertainment boxes at home), and that image dissolved. It took a few years - and the fictional tropes of actually sapient computers turning into Frankenstein's monsters lives on - but the bigger chunk of that, the misunderstanding of computers in the most fundamental of ways, was over. But by that time, they'd already become so useful that there was money to be made - particularly in software - so.

Secondly, there's no war.

And it was a war. It really was. There was a generalised war against power and central authority going on in the culture in the 70s, and this was a critical part of the context. Sure, it was in its waning moments in a lot of areas, but in this particular crazed section of nerd culture - lagging popular culture here, really - it was absolutely going strong. And as... quaint as it may seem now, part of it was absolutely going to war against IBM and against the elevated-floor refrigerated-white-room centrally-controlled data-priest-blockaded batch-processing punch-card era it represented.

(That's one of the reasons the IBM PC was so hated by so many already in microcomputers, and so many of the new companies threw themselves at it so hard, trying to knock it over - and also why IBM's "legitimisation" of the microcomputer mattered so much, as well. IBM's PC was honestly nothing much - but it carried the magic three letters of capital-A Authority that turned business perception of microcomputer from "toy" to "tool." If you're at war with that capital-A Authority, then the hatework generated? Enormous. You'd've caught some of that in the Netscape-Microsoft war of the late 90s, but the key difference there is that Microsoft wanted to own the web, whereas most of IBM hated the PC, and saw it as a way to leverage people back to the mainframe. Which in turn is also why they eventually lost.)

lol, war stories. But back on topic: both of these cultural factors are huge motivators. Huge. And they both leveraged the batshit insane amounts of work needed to turn those tiny chips into machines actually worth using. And, from here, it looks like that's mostly missing right now, with both the robotics and fabricator hackers. You see hints of the first in singularity fiction - actually, I hope you do - but that could be wishful thinking on my part.

And of course you're totally right about the scaling issues not helping. It is just harder, like AIs have been just much harder. In both cases I kind of like to hope the tools aren't there yet, but will be, and then things will get awesome. I hope they do.

[identity profile] randomdreams.livejournal.com 2011-06-14 12:02 am (UTC)(link)
Weirdly, Rigel and I were discussing this today.
My friends who have 3D printers, both commercial and opensource, spend a *lot* of time screwing around trying to get them to print right, and throwing away a lot of didn't-work-quite-right parts.
It sure looks a lot like people messing around with ZX80's.
But people who have spent the time -- and I'm guessing it's in the hundreds of hours -- to get their systems working the way they want, have started doing some really cool stuff: ooh, the mounting plate for the fairing on the recumbent broke! print a new one! (http://softsolder.com/2011/03/15/tour-easy-zzipper-fairing-upper-mount-plates/). Oooh, the micrometer handle broke! print a new one! (http://softsolder.com/2011/05/27/thing-o-matic-caliper-repair-perfection/).

I can see a time when it's going to be the same feeling you got when you bought the MIG: suddenly you look at the whole world in a different way, disassembling parts into their printable components. It doesn't look like it's that way quite yet, but we're reeeeaaaallllyyyy close.
Since 3d printing is unlikely to be Moore's Law-compliant, we'll have a lot more time to get into it, once it starts swinging more heavily.

[identity profile] randomdreams.livejournal.com 2011-06-16 04:05 am (UTC)(link)
It occurred to me yesterday, after messing about with the CNC mill with a marking laser in the chuck, that the mill is basically a positioning device and there's nothing aside from good sense stopping me putting a Thing-o-matic Mk6 extruder head on it in place of the spindle and doing some 3D printing.

But that just raises the same question: then what?