If there was a big crisis, We could retreat back to 1930 level of technology with relative ease in fact, maybe even later. That gives us radio, telephones, telegraphy, et al. All that tech can be made in small workshops.
Those machines required a large number of precision stampings and small machined parts. Lots of custom tools and dies. Tooling up to make all those was a big job. Once they had all the tooling and plant, the Model 15 was cheap to make, despite its complexity. Teletype tried making a cheaper, simpler Model 26, but it cost more to make than the Model 15 because the manufacturing process for the heavy-duty machine was working so well.
From the beginning, most Teletypes were rented, not sold, and came with maintenance included. So the machines had to be reliable. The Model 15 seems way overdesigned. Flat parts were both hardened and Parkerized, a caustic process that deposits a rust-resistant oxide and turns metal grey. (Sometimes used for guns.) Fine-thread screws. Lock washers on every screw. Insulation that won't age or crack. Plus, the keyboard, typing unit, and motor are all very easily replaceable; they even disconnect electrically and slide out. The screws and round parts are all custom. They must have had a huge number of automatic screw machines turning them out. All round parts fit precisely.
This was not made in a small workshop. More like a factory that covered a city block in Chicago.
WE Hawthorne Works in Chicago was similarly large, and to build in quantity you'd need it to be just as large now, but to build a couple hundred strowger switches a month? Not as large.
The hardest part of bootstrapping to 1930 is making copper wire - doubly so for wire suitable for electrical coils.
A good chunk of hobbyist-machinist YouTube is essentially just people bootstrapping themselves into a complete shop's worth of tools starting from three or four primitives, and it's kind of wonderful to watch.
(I'm always kind of disappointed whenever one of these hobbyist folks get a CNC mill or something, because it throws off that bootstrapping feeling. I know it was never their goal, but they were coincidentally pulling it off until that point.)
Another thing is audio amplification can be done without tubes or transistors, you can use a mechanical amplifier which is literally a transducer that is mechanically coupled to a carbon microphone, and a local power source (battery) making it possible to build analog phone lines that run several hundred miles without tube type amps.
(And then the same is true of everything after that, because transistors require photolithography, and photolithography requires good lenses.)
Of course, if you’re unlocking your tech-tree in a breadth-first direction, you’ll already have the glassworking primitives around due to them being common prerequisites for refining metal ore :)
I'd like to point out another interesting 18-part series of videos about prototyping (in his words "a short course on how to build stuff" :)), by engineer/inventor/professor Dan Gelbart: [1]
[0] https://archive.org/details/GingeryMetalWorkingShop/mode/2up