But those giants of the 1970s fell victim to what's now widely called disruptive innovation, a concept popularized by the late Harvard professor Clayton Christensen. It's what happens when a product starts out at the bottom of an existing marketplace and eventually moves up in value to replace the established competitors.
The makers of supercomputers and mainframes, even IBM, were originally dismissive when a then-small maker of memory chips called Intel Corp. built a "computer-on-a-chip" for a Japanese adding-machine company.
I used CDC computers from about 1983 to 1992, NOS and NOS/VE operating systems. Truly awful.
CDC, at least, did not "fall victim" to anything. They decided they were the best, they battened down the hatches and refused to change even one little bit. They actively refused to change with the times, to incorporate any new idea.
The HP PA-RISC chip was a huge disruption when it came out. It led to cheap, fast workstations which killed the mini-super vendors (Convex, Stardent, ...). These vendors had already damaged Cray and the other large-scale vendors. And most importantly, the HP chip could run Windows in SoftPC faster than Intel's 486 could run it in hardware. I think it was 1992 or thereabouts, and it was a mess.
Sh*t changes, and perhaps the most interesting lesson out of all that is that software rules and hardware is merely their complementary product.
I remember getting a trip home with a mainframe programmer in the late 1970's. The entire time he spent telling how limited micro-computers were and how much programming knowledge he had.
I have doubts he could handle the machines that started going into businesses in the 1980's.
The makers of supercomputers and mainframes, even IBM, were originally dismissive when a then-small maker of memory chips called Intel Corp. built a "computer-on-a-chip" for a Japanese adding-machine company.