Reasonable risk-managers only go so far. There's always the "big asteroid that goes undetected" that lands on your building during a big in-person meeting tha thas most of your company's key talent.

My recent backups might be infected, but my "day of compromise minus one" backups won't.

Even if my recent backups are infected, they are likely to not be ransomware-encrypted, which means they are still useful to me.

Technically, this is very hard to do. It's much easier to set things up so there should be no operational way to "delete" or "modify" existing records without it being obvious that something out of the ordinary is going on

With the right level of access, there will be a way to copy everything from the existing media EXCEPT what you want deleted to new media. As long as this is easy to detect (say, CCTV recordings showing someone entered the server room, downed the server, removed the write-once media, used a magic box* to copy only what he wants to copy, then replacing the old media with the new media), that's going to be a deterrent to unauthorized record-deletion.

* how the magic box works is left as an exercise to the student, but for planning purposes, assume such a box exists until proven otherwise

Microsoft is stalking you. Again.

Imagine it's 1950s or earlier. You run a business that lives or dies by paper records, such as an insurance company, land office, or something similar.

Your office burns down, taking all the data with it. You don't have off-site backups (microfilm, carbon copies, or what-not). Thankfully the fire was after-hours and nobody was hurt.

Your business is probably toast, figuratively and literally. At best, you are insured and will be able to start over from scratch, but your existing customers might prefer to start over with a company that knew how to keep backups rather than continue working with you.

Re:What do you mean by "possible?":

The oldest monster6502.com entry in archive.org is from 2016.

If true, this is the modern-day version of a Letter of Marque, with the slight (cough cough) difference that the United States and China are neither technically at war (like N. and S. Korea) nor actually shooting at each other (like the various non-declared wars/hostilities the US has been involved in after WW2).

If the raw materials are all on one continent, the end users are on another, and the finished product is less massive than the raw materials, it's going to require less shipping to build it where the raw materials are then ship it to the customer.

Also, what about things like coffee, that simply don't grow everywhere they are consumed?

If I read this right, ancient seafaring peoples had the same problem, just on a much smaller scale.

"No, you can do the equivalent work of earning a _DEGREE_ (minus the non-technical classes) by on-the-job experience or self-study without earning a degree and still be a good developer."

Sorry about that.

My keyboard dropped out of college shortly before graduation. That's my story and I'm sticking to it.

No, you can do the equivalent work of earning a (minus the non-technical classes) by on-the-job experience or self-study without earning a degree and still be a good developer.

"4 year CS degree or equivalent education or experience or combination thereof" is a much more rational hiring criteria than "4 year degree."

We may not have a "trades system" but we do have privately-run certification systems that are separate from university degrees.

Think of an entry-level CompTIA/Microsoft/Cisco/OtherMajorBrand certification as apprentice-level. You can move up from there and specialize without needing to go to a college or university.

... are belong to us

Fail.

For $299 I want as close to an original-part-by-part reproduction, not an FPGA cheat.

For an FPGA cheat, just publish everything open-source and sell a completed product for those that don't want to DIY.

I'm willing to compromise a bit on the FPGA front for individual chips that are unobtainum, have no modern substitute, and not cost-effective to recreate in small-ish quantities. However, the substitute chip should be pin-compatible and of course timing-and-everything-else compatible with the original. In other words, if I had a bin of original parts and a bin of the FPGA substitutes, I should be able to use either one in a rebuild without noticing the difference. The motherboard, connections, and feasible-to-reproduce parts should be faithful reproductions of the original.

I'm hope this product is successful, but it's just not for me.