China has some of the lowest labor costs on the planet, and they are automating basic tasks like mad.
The reason being is they intend to keep costs low, and the only way to do so is to create a nation of billions of semi-skilled robot tenders. The automation can be set up quickly and efficiently by a group of skilled technicians, tooled to a task and set in motion. It is faster cheaper and better to do this than risk having people gain skills and price themselves out of the market like other countries have.
It's happening now, I'm sitting back and watching it from where I sit. We were counting on China eventually pricing itself to death like other countries did, but they seem to be making a serious effort to prevent it from happening.
So the only way left forward for American Manufacturing is ultra-automation of even basic tasks. We simply cannot compete any other way. Where I am sitting I get a lot of companies calling up looking for repair parts for old automation machines, fixtures and production tooling. In the 1990s a lot of packaging and asssistive productive equipment was idled as companies scrambled to bolster workforces in order to gather as many tax-credits as possible. This was about the time that incentives were offered wholesale if companies would bulk-hire people with X demographic qualifications or Y profile scores... The logic was create a false turnover, report that you were "growing" your workforce with each turnover and pay little to no tax because of the credits.
Lines were retooled back to manual production, fixtures removed, packing machines idled and large numbers of temps doing physical manual labor were employed. In many ways 1988 to now represents a tremendous step BACKWARDS in terms of productivity and automation. I've looked back through old publications and seen examples of automation and labor-savings that just don't exist any more. When I mentioned them to leadership and outlined the benifits of not having a tremendous turnover dragging down productivity and quality I was told "then you wouldn't have a job -- " Well I ended up without a job anyway in 90 days when it was time to roll over the workforce...
One of the worst examples of the removal of automation was a plastic bag factory in NY I worked at. They made a tremendous number of Hefty and Kordite branded products, originally the company was Mobile Chemical. They sold off the division to another company... company decided to cash in on the credit-scam tore out all the automated packing and boxing systems. They did away with the automated batch-weight material formulation/mixing system and discarded the automated warehousing system. All so they could hire manual labor in bulk so they could claim tax credits.
Building has been empty for well over a decade and a half now. It was sad to see them go, but the price of material and energy spiked and they were unable to compete in the market. All that production was consolidated and sent to plants that were automated and that was that.
Now at the small scale, the "home-shop" and "micro-enterprise" level many of us represent, automation of certain tasks does make sense. I have an injection mold machine for production of batches of parts... My CNC mill can do many operations that I'd have to do by hand on a drill-press.. I have a small Fanuc robot I can program to do pick-and-place... it's a simple matter to print a fixture, program the robot, make an EOA gripper or tool out of bits and pieces (air cylinders, valves, printed parts) and set it loose.
We just completed the production of 250 gripper/positioned for a bottling line that was rebuilt and restored to operation. The company in question shut down the capper section of the line and hired 50 temp workers... it's no longer viable to have 50 people manually cap soda bottles so they rebuilt the line. We used the robot to pick up the parts from a stack, drop them into a fixture, the mill verified all six holes (drill cycle) and then put them into another stack for me. While it did that I worked on other projects.
It can be done, it's a mindset, a skillset and it requires imagination. Audrino-boards, ras-pi modules, cheap abundant sensors, inexpensive simple robotic kits... 2020 profile and an idea. That's all it takes. You can program up a fairly complex sensor-monitored action sequence these days using these boards. In many ways they are just as flexible as a full-scale HMI-equipped PLC system. If not more so.
If you have read this far, what I am saying is automation can be done on the small-shop scale and it can be done far cheaper than you imagine. If you can knock together stuff like your keyboard product, you are already are already half way there.
If you only deal with the "reality of numbers" and text-book graphs and examples, you will never see a solution.