>>11974756Why are you here asking? There answer is obvious. In order to make more performant stuff, you need to improve your control over things such as...
- error
- uncertainty
- accuracy
- precision
- margin
Each is an entire conversation, but it all boils down to...
- better vertical integration (more elegant but usually less computationally efficient models), which grants you capabilities in the vein of DARPA's so-called SSITH; this grants you a higher ceiling for the most complex things you can develop in an SSITH style
- a lower floor at the bottom meaning for example making nanowires without relying on macroscopic emergent phenomena making them for you (like shooting a laser pulse at a micrometer gold wire to make a 10-nanometer gold wire "bridge"), again less manufacturing-efficient but easier to represent logically (more elegant, easier to wrap your head around the system) in your SSITH system
The practical consequence is...
- Learn how to rely more on harder logic to speed up your work flow, and re-allocate your saved time toward better documenting the intuitionistic aspects of your discovery, invention, maintenance and lifecycle management, etc processes for some system
- Building ever more precise machinery.
The former is self-explanatory, and the latter boils down to making ever-more-precise CNC machines without hand-designing (but with hand-tuning/maintaining) them, which is to say instead of hand-programming their firmware, your budding SSITH design meta-process should make and load the firmware for you.
At this point, you would get stuck, but it again boils down to either making your own mcus or else relying on fpgas on the one hand, and on the other hand, drilling down in the world of macroscopic manufacturing until you are playing around with soliton waves for microscopy and holographic rulers for feedback.
Then you can mass-experiment in MEMS and all of the sudden the world is your oyster.
Obviously you cannot buy this stuff online.
