ARM charts out the path to printed plastic chips
semiaccurate.com
semiaccurate.com
What exactly are they printing ? Are the transistors actually made of plastic instead of doped silicon ? If so what feature sizes are possible with this technology ? How does the manufacturing process compare with Si ? etc...
What you can imagine out of this technology is a moore's law type reduction in feature size. Not too far off you could imagine a box where you could print arbitrary chips on a whim.
What it is is orders of magnitude cheaper and faster production of complex microchips.
This sounds like it could become very important. Do you know of any sites that do a better job of explaining this technology ?
Specifically, it allows orders of magnitude cheaper and faster one-off chips. The manufacturing method they use will always lose against current methods when making a lot of chips in bulk, but right now if you want an ASIC, the first chip costs the same as the next million, and this is something they can change.
Red flag when a guy writing about bleeding edge transistor technology doesn't understand the main point of Moore's law - exponential growth in density, which is usually converted to a log scale making it appear linear in nature.
Transistors are dirt cheap in volume.
In reality, transistors have a megabuck+ NRE (non-recurring expense) that prevents us from making effective use of the global oversupply of transistor manufacturing capacity.
So, just like plastic injection molding is dirt cheap in large volume, but 3D printed plastic is way better for small volume; silicon transistor manufacturing is dirt cheap is large volume, but printed transistor manufacturing would be good for small volume.
The plastic process allows you to make masks very cheaply (I believe on the order of $5k, as opposed to $millions), so low-volume production is more practical.