The Microfactory: A machine shop in a box | Hacker News Reader

The Microfactory: A machine shop in a box

kickstarter.com

31 pointsbshanks18 comments

In the interests of transparency, please note that one of the four founders didn't complete fulfillment on his last maker-oriented Kickstarter: https://www.kickstarter.com/projects/1738994529/kikori-open-...

It's amazing to me that Kickstarter allows delinquent creators to start another project.

I don't know how they'd track it when it's not the same account.

I pledged to the Kikori Kickstarter (which, btw, means I've got some bias in this although I wasn't up for any rewards) so I knew who the guy was, and I happened to notice that he was involved in this one. Even a human reviewer wouldn't have caught that. Well, you could do simple name matching but I don't think that's good enough -- compare this to the no-fly list, for example.

So I'm OK with word of mouth being the governor here. It's also a good indicator that you should research before you pledge. I do a quick web search if I'm backing something from someone I don't know. It helps set my expectations.

Kickstarter campaigns do have to be submitted for approval, so theoretically there is a human reviewer. I'd assume step 1 would be "Google the names of the project founders".

Sure, but then you have the false positive problem that plagues the no-fly list or any other string based name matching system. It'd work for Judah Sher since that's a fairly uncommon name (his Kickstarter is in the top ten results). It wouldn't work for plenty of other people.

That thing looks amazing, but probably before its time, $5k is a lot to throw down on a preorder for a machine. I would happily put down that if the thing was already available though... perhaps they should explore more traditional funding routes to get production started?

You can get a milling machine + lathe + a bunch of tooling for way less than that, and have plenty left over to make it CNC if you so desire.

http://littlemachineshop.com/

Also: http://www.mini-lathe.com/

Not to mention that LittleMachineShop sells complete proven CNC rigs for about the same price with no risk, and which are obviously much more rigid as well.

yes, the kickstarter is expired, but https://twitter.com/Mebotic indicates that the project is still alive.

Hey guys, just saw this thread. Yes the project is very much still alive. we are making a less expensive version of the machine which we hope to have up on Kickstarter in the near future. In the interim we are having a ton of fun posting projects we've been making with the machine which you can find on the bottom of our homepage www.mebotics.com

THANKS FOR THE SUPPORT!!!

How much longer till a machine can produce an identical, working and complete copy using only raw materials and a blueprint?

After we discover a way to make microchips in factories that cost $1k instead of $1G, and deal with the revolution in economics and computing that follows it... Then we can think about machines that reproduce themselves.

As chip technology advances, does any of it make it substantially easier and cheaper to make old, slow chips? In other words, while massive factories are clearly the way to go for current technology and they let you pump out cheap chips in massive quantities for ridiculously low per-unit prices, what if you wanted to, say, fab an 8086 today? If you jump a few decades back, can you make those chips individually today through some reasonable process, or is the whole business just inherently large-scale even after all this time?

An automated chip factory intended for use in a self-replicating machine is only useful if it can produce chips from unprocessed ore. You can make your job a lot easier if you design it to just run on silicon boules-- but those are almost as expensive per kilogram as finished packaged chips...

So you need a furnace for refining sand, a chemical factory to produce the dopants, lithography equipment, clean rooms, ultraprecise steppers...

Really, the problem here is the hidden complexity. It's not like you can look an obsolete chip foundry, and say, "Okay, so this facility is worth $100 million, and is about an acre of equipment, how do we miniaturize it?"

That single foundry is the tip of a $100 billion pyramid. You can't just replicate the foundry, you have to replicate all the factories that built the stuff that went into the foundry.

How this usually works is someone comes up with an entirely new way of accomplishing the task. Someone young or from outside the industry who doesn't know what they're trying to accomplish is impossible.

When I look at how fast the 3D printer industry is going you never know. If you can currently print a house using concrete and in the near future print out a kidney who is to say you won't be able to make an 8086?

http://www.wimp.com/printerhouse/

http://www.livescience.com/41480-3d-printed-kidneys-take-sma...

Programmers typically don't have very good intuitions regarding the pace of change in other industries.

Quoting myself (http://c1qfxugcgy0.tumblr.com/post/31187427192/the-enduring-...):

  Not every technology follows the price curve of 
  computers— computing is the bizarre outlier. If you had 
  predicted in 1970 that computers would be tens of 
  millions of times cheaper per FLOP in the future, you 
  would be a visionary, but if you predicted that jet 
  airliners would cost $10, fit in your pocket, and fly at 
  Mach 100, then you would be an idiot.

Some things are just hard. Not everything gets 10X better per year for decades and decades. Most industries hit hard physical limitations long ago, and are improving either very slowly, or not at all.

3D printing is a particularly bad example of revolutionary change: it was invented by industry insiders, for decades was performed by exorbitantly expensive machines, and is mostly hyped based on the term "3D printing." Actual FDM machines are much less exciting, and aren't terribly useful, commercially.

Depends on what you'd call a "machine", and how "raw" your raw materials are.

Bob Freitas wrote an open-sourced book on this a decade ago: http://www.molecularassembler.com/KSRM.htm

For simple tinkertoy-ish machines, floating in a sea of parts, and sitting on a shaker table, (to provide power) it's already been done.

For anything with a microchip in it, full closure would require machine-phase nanotechnology. Partial replication should be possible with conventional technology, (a "clanking replicator" http://en.wikipedia.org/wiki/Clanking_replicator ) if you supply the replicators with "vitamin" parts they can't produce themselves. (Chips, stepper motors, anything made of rare minerals)

A lot of work was put into this for the Advanced Automation for Space Missions study in 1980, a proposal for a solar panel factory on the Moon that would build more solar panel factories: http://www.islandone.org/MMSG/aasm/

Reply on news.ycombinator.com