4 ms·

There's a few ways to look at this: In the long term (30-50 years), "replicator" technologies could appear, and then game over there's some major disruption co

There's a few ways to look at this: In the long term (30-50 years), "replicator" technologies could appear, and then game over there's some major disruption coming. However, for the type of machines today, better democratization means easier prototyping, which combines with hackerspaces / TechShop, open hardware, etc. are already allowing hardware startups to emerge - a movement that's gaining steam rapidly. For the mid-term, cheaper prints comparable cost/materials-wise with injection molding, this could still do some pretty amazing stuff with flexible production, but we'll still need assembly lines with chips, etc. Maybe if we can get conductive inks to be responsive enough for digital electronics, we could print circuit board inside and just pick-n-place the ICs. Also, by this time we may have a more modular approach to hardware where only a few thousand components are needed to make 99% of devices - then Kinkos-scale production would be possible. Timeframe: 10-20 years? What do you guys think?

se85 14y ago
That would be great, but I think the time frame isn't realistic unless we already have the technology to do this but it is not out of the lab yet so to speak. If we do, then perhaps it is possible to make the manufacturing processes simple enough within this time frame to bring the costs down to the point where these types of companies would be commercially feasible.
- nickpinkston 14y ago
  Can you elaborate on your answer and time frame question?
ippisl 14y ago
The whole electronics supply chain is optimized towards volume manufacturing. There's a second supply chain for small volumes, but it's expensive: boards instead of chips or more expensive chips and boards(due to volume breaks), complex development, testing. The assembly costs are relatively minor. I think the first step for the kinko[1] vision would be to largely automate the development and test costs. Maybe arduino is one path to this. [1]I think it would make more sense as a centralized source who prints designs coupled with rapid shipping, not as a kinko.
- nickpinkston 14y ago
  Yea, it's true on supply but this is a structural problem (i.e. minimum stocking sizes at 4K units, etc.) that would change if the demand was there. Agree with [1] - that's partially already here with ProtoMold, Sunstone Circuits, Ponoko, etc. - centralized makes sense when you have economies of scale, but my comment was that this scale can actually be driven lower via technology. I think that development / test costs are a big deal, but maybe the "automate" part is more in the EDA sense of the term - than "sudo make me a sandwich" where your design is compiled directly.

se85 14y ago

That would be great, but I think the time frame isn't realistic unless we already have the technology to do this but it is not out of the lab yet so to speak. If we do, then perhaps it is possible to make the manufacturing processes simple enough within this time frame to bring the costs down to the point where these types of companies would be commercially feasible.

nickpinkston 14y ago
Can you elaborate on your answer and time frame question?

ippisl 14y ago

The whole electronics supply chain is optimized towards volume manufacturing. There's a second supply chain for small volumes, but it's expensive: boards instead of chips or more expensive chips and boards(due to volume breaks), complex development, testing. The assembly costs are relatively minor. I think the first step for the kinko[1] vision would be to largely automate the development and test costs. Maybe arduino is one path to this. [1]I think it would make more sense as a centralized source who prints designs coupled with rapid shipping, not as a kinko.

nickpinkston 14y ago
Yea, it's true on supply but this is a structural problem (i.e. minimum stocking sizes at 4K units, etc.) that would change if the demand was there. Agree with [1] - that's partially already here with ProtoMold, Sunstone Circuits, Ponoko, etc. - centralized makes sense when you have economies of scale, but my comment was that this scale can actually be driven lower via technology. I think that development / test costs are a big deal, but maybe the "automate" part is more in the EDA sense of the term - than "sudo make me a sandwich" where your design is compiled directly.