Slashdot Mirror
Open Source Self-Replicating Robot
Josilot writes "CNN.Com is running an article about a new self-replicating robot named RepRap. From the article: 'A revolutionary machine that can copy itself and manufacture everyday objects quickly and cheaply could transform industry in the developing world, according to its creator.' One part of the article that I think many slashdot readers will find interesting is near the bottom: 'To encourage that development, Bowyer plans to make the design of the RepRap available online and free to use, in the same way as open source software such as the Linux operating system or Mozilla's Firefox browser.' Is robotics the next big field for open source?"
I want a box in my kitchen that makes synth-protein from scratch!!
"I had another dream the other day about music critics. They were small and rodent-like with padlocked ears..."
One part of the article that I think many slashdot readers will find interesting is near the bottom:
New here, aren't you?
Have we contacted the Asgard? Smells like trouble to me!
mazevedo
I've got a better open source self-replicating robot: Herpes virus.
BTM
That was the turning point of my life--I went from negative zero to positive zero.
new toys .. more new toys .. wait that isn't a toy, why are they eating my house.. humm i don't think they like me
'...if only "Jumping to a Conclusion" was an event in the Olympics.'
If it wasn't, would it's own existence violate the DMCA?
Stasis is death. Embrace change.
Yes, of course robotics is the next big field for open source! The FOSS model works pretty damn well, it would be (in some people's opinions) selfish not to apply it to other aspscts of technology and life in general.
North American natives did something similar to open source by sharing their ideas, methods, and beliefs with the Europeans that came to North America, and the Europeans gave them the advantage of metal pots and pans. Basic open source right there. Now we have North American society, home to the most powerful country on the planet.
It may sound cheap to you. But consider this:
A laser need not necessarily put out much power to fuse even metal, if the metal powder is already close to melting point to begin with. An existing prototyping system uses this approach to create customized metal objects.
BTM
That was the turning point of my life--I went from negative zero to positive zero.
RTFA, it's not a dupe. Looking at the picture in the article it dosn't look like a bunch of cubes moving themselfs around as in the first case. It looks more complicated now.
Cheers,
RoadkillBunny
The article is frustratingly vague. It sounds to me like the robot doesn't replicate itself but rather that Dr. Adrian Bowyer has created some type of system for replicating robot chasis. The picture clearly shows a plastic 'bot with attached motors, wiring, batteries, etc. From the information that the article gives, it seems like a human is still needed for the final construction. I wouldn't consider this self replicating because it is not autonomous.
We've had 3D printers for quite a while now which basically form shapes by laying down a thin layer of sand-like or metallic powder, followed by a thin layer of glue, etc. You then use compressed air to blow away the sand layers which don't have glue and voila... a 3D shape and quite sturdy. You can make some parts which are impossible using other methods.
However, I missed the part in the press release, er... story where they are self-assembling. Sure, you can have a machine feed in a design and print something out, but what about assembly? Yes it can print circuits, but does this thing add motors, insert batteries, or plug its power into the wall? And will it feed the newly created copy with the source of materials, etc. it needs to make another copy? Let me know when we get a machine which can create an copy of itself and, without any human intervention, that just-created copy makes another copy.
Want to improve your Karma? Instead of "Post Anonymously", try the "Post Humously" option.
It's apparently a robot that can make circuit boards, and that's it. There are about 50 million steps involved in making itself, this can do one of them.
Thx media hype, call me when something interesting happens.
http://reprap.org/
the blog is cool too
http://reprap.blogspot.com/
This would make a better type of bot wars, building their weapons with available materials and blasting each other with them heh
With the billions of inhabitants this planet has, are we concerned with robot workers?
ZP
We only can learn from our mistakes.
At least they aren't powered by human blood.
--
make install -not war
Generation 33 = One for everybody on Earth, plus about 10% overage. Let's stop there and start making stuff with them.
In fact, what will probably happen is everyone who gets one going will make a copy or two (on the average) for people near and dear (for average values of dear). Then they will turn them to making other stuff. That means it will spread much more slowly than exponential growth. A slower growth rate is good from a control standpoint, bad if you are waiting for them to spread to your area and lack the skills to jump-start the process.
What's neat is having someone make their own replicator simultaniously teaches them how to use their copy for making other stuff, unlike sex.
Who is John Cabal?
That's silly. Fabrication speed is still an issue even if they are self replicating.
If I need an entire warehouse full of self replicated robots to fashion a plastic spoon in under a week then it would not be practical.
Alternatively, if they can only manufacture things solo (especially small things.) then their ability to replicate does not enter into it.
Lastly the bots present an overhead. Their raw materials must be paid for and the bots must be powered. (It looks like it runs on a few D cells. If I had a warehouse full of them I'd need an army of people just changing batteries for me.)
I'm not putting down this invention, (Though I don't think it's as far along as CNN would have us believe.) I'm just pointing out that self-replication does not necessarily translate into manufacturing efficiency or free wealth for all.
It's about as self-replicating as a machine that connects to the web via its ethernet port, places an order for parts here, waits until the UPS web site says the parts have arrived and then emails its owner to tell it to assemble the parts sitting in the box on the front doormat.
Doesn't it make you feel good to know that our freedoms are protected by politicans, lawyers and journalists.
Man, I can't wait to download the latest PowerMac G7 from Kazaa, 2 days before Apple releases it...
I don't care if it's 90,000 hectares. That lake was not my doing.
Generation 31 = robots use up the last of the available IPv4 addresses and turn on their masters, subjugating humanity and forcing it to adopt IPv6 at gunpoint.
The horror!
I don't care if it's 90,000 hectares. That lake was not my doing.
First, it's been on Slashdot before.
Second, the "self replication", as the paper puts it, consists of making a baseplate in an fused-deposition modelling machine and depositing some metal on it, to make a very low density circuit board. That's all. This is an expensive way to make a crappy single-sided PC board. It's a giant step backwards from the photoetching processes used now to make both PCs and ICs. Not only is it nowhere near "self replication", it's not even a step in that direction.
Finally, the web site has the tag line "Wealth without money...", which sounds like something from a stock scam.
Actually, the cutting edge work in this area is not fake "self-replication", but using deposition-type fabricators to make 3D objects with complicated internal structure, including combinations of flexible and rigid materials and moving parts.
There's a fundamental misunderstanding about manufacturing that pervades enthusiasts for computer-controlled one-off manufacturing. It's that most manufactured goods are made by some process that involves a "master" or "mould" or "die", and that those processes are incredibly cheap. There are about a hundred such processes in common use, from injection moulding to photolithography. And they work quite well. That's what you're competing with. Making single parts in bulk just isn't that expensive.
yes, it's a hodge-podge, make-shift, kludgey mess.
but make no mistake. no matter how imperfect, hobbled, or inelegant - the first von Neumann machine will start an exponential avalanche.
welcome the singularity.
K.
I certainly hope they've figured this one out, but if not, I'll repeat it here: Robots do not have to carry their "brains" on board. Just have them "download" their instructions from one central hub, which can then be easily re-programmed/de-bugged without recalling every unit.
Seriously, don't *anybody* even let this one slip to the US military. Shhhhh! If you don't want to see the entire planet burned black by self-replicating tanks and bombers, don't even tell the US Gov that these things exist. Especially not if you have oil.
Everyone pointing out that mass production is far more efficent and that a factory of these could never compete.
Thats not the point. These things are not designed to compete at that level. A one step printing process like this will never compete with mass production methods for speed. What it trumps that process in is versitility. How much does a prodection line for a screw cost ? How much does it cost to create a new screw design to implement ? How much to switch between certain templates ? How many must be made and sold to make the process profitable.
In short mass production relies on economies of scale. Makes lots of goods far cheaper than they would be otherwise but at the same time it sort of forces us all into a one size fits all world where the only things that get made are those with a large enough mass demand to support the enourmous intial investment in establishing such a process. Cheap is a very relative term when speaking of mass production. See if you think any of the numbers involved in setting up a first run of an item are 'cheap'. This creates a staggering bar to market entry in many fields.
What they are not good at is adapting to needs and they require enourmous amounts of stock to be made and shipped before demand is established. Distributed production like this would do a great deal to elimnate overstock. It could potentially lower the bar to market entry in any number of areas. For example lets say manufacturing shifted from highly specific highly concentrated mass production to highly disperesed general construction. In otherwords demands to keep such a process running would run the gamut of production needs rather than rely upon one specific need. In other words more smaller factories capable of producing A-Z instead of just differnt sizes of A.
For example If someone could create a rare auto parts fabricator that worked cheaper than machine shop rates for custom replication they would make a fortune. Demand is there for such a thing but no one items demand is large enough in most cases to permit someone to make money setting up a mass production line for it.
Also imagine the new frontiers opened up to product hackers if they could alter the design specs acording to their whims rather than be stuck with what is profitable for a mass production run.
I don't ask you to be me. I only ask you not expect me to be you.
The here is the googled website: http://reprap.org/ I have always thought replicating machines would be cool (it's possible as can be seen from the "two-legged existence theorem"). Using a 3D printer sounds like the way to go about this today. However, I would think that whatever it is you are crafting with a machine has to necessarily be at a lower mechanical tolerance to the machine itself, so over several generations, the precision falls catastrophically. Some form of recovery or repair is needed (as in DNA). I have always thought we were a long way off from this sorta thing ... ...
The practical solution outlined or implied here seems workable in contrast - use your machine to make various parts, and have a human assemble them together. It was mentioned in
http://www.blog.speculist.com/archives/000293.html
that additional off the shelf spare parts might be needed. This certainly provides a practical trade-off for an almost self-replicating machine. Nice! Hmm
"We'll be watching your future career with great interest." (Senator Palpatine)
i agree that the "self-replication" stuff is b.s.
on the other hand, as someone who photo-etches my own circuit boards and finds the process quite archaic and annoying, it would be really excellent to have something that would squirt out copper/alloy traces and drill the holes for me. like a little ink-jet printer for pcb's.
There's a fundamental misunderstanding about manufacturing that pervades enthusiasts for computer-controlled one-off manufacturing. It's that most manufactured goods are made by some process that involves a "master" or "mould" or "die", and that those processes are incredibly cheap.
i'm confused by what you wrote, you say that the belief that most manufactured goods use an inexpensive mould/die approach is a fundamental misunderstanding? ok, i've read your paragraph several times, and i think maybe you actually meant to say that these enthusiasts don't understand the reality, which is that most manufactured goods use an inexpensive mould/die approach. is that correct?
in that case, i would just want to counter by saying that the process of creating the masters or moulds, eg. the tooling for an injection mould, has traditionally been a relatively large expense.
if you can put the ability in peoples' hands to create inexpensive injection moulds and other masters, i.e. by providing a cheap/open milling machine design, along with sophisticated free software for 3D design and automatic machine control, you open a lot of doors for small businesses - even if the product will ultimately be mass-produced in a traditional factory mould/die process. you reduce the up-front expenses for highly-skilled design engineers, machinists, and pattern-makers, by allowing a more trial-and-error approach to initial design and testing.
ultimately that will mean a lot of really crappy design - in the same way that laser-printers and desktop publishing software fostered a lot of really crappy print design. but at the same time... it was a big paradigm change, and the design and printing industry will never be the same.
this "reprap" project is unfortunately full of the typical fantasy/hype of some student projects. but if you can see past the hyperbole, you can see that it is, if not creating, then at least responding to some quite interesting steps forward.