As the article belatedly mentions, the ~40 GB automated tape library for HARVEST was the bee's knees... and possibly the inspiration for "harvest now, encrypt later".
PS – Primary sources for the IBM 7950/HARVEST on Bitsavers, excellent as always:
One of the big features of the Harvest was the so called POPCOUNT which was supposedly added the machine at the NSA's request. (This is the number of bits in a word that are on.) I've always wondered how this might be used. Some suggest computing the Hamming distance. Others suggest some kind of statistical test like chi-squared. But are there other more algebraic solutions that can benefit from it?
Every time I read an article like this, I get depressed about all the hours of my life I've spent waiting for computers. I'm 64 now, and I've been using computers since I was 14. (Batch mode, using punched cards on an IBM 370 was the first computer I programmed, while I was in high school.)
They had vertical integration and made bank. Then each layer got more complex as computing became mainstream, and IBM got out-competed on everyone of them by specialized companies. And then they sold them off. The only layer that remains is the most strategic: the customer relationship.
In some areas IBM still holds if not a technological advantage at least the first among equals position, they just aren't areas that are flashy or generally known outside of folks like us - Mainframes and related technologies for example, the IBM Z series is still a huge player there.
Sometime in the 90s they decided that becoming a services and consultancy provider was more important than being involved in large-scale commodity hardware or really anything else.
Over three decades they've exited commodity hardware (x-series and thinkpads), retreated from a lot of software stuff and generally gone as hard as they can for large-scale business contracts. They seem to keep a decent amount of cutting-edge hardware research going, Power is still broadly competitive with x86 AFAICT, but the volumes have got to be pretty low these days and I half suspect that some of it is just a "Hey look at us we're still here" play rather than actually feeding into products, because services are the profit-centre.
They're still huge and have a lot of money, and they own Redhat. But IMHO it's in a long, slow decline, and it has stepped back from the public eye and most public-facing markets.
Context - I've worked for them in a few places. Obvs my opinons are my own, especially as I've been 'out' for over a decade at this point.
They royally screwed up by signing an exclusive deal with the back-then no-name Microsoft company. IBM created the 3 trillion behemoth that Microsoft is while killing themselves in the process.
It's really sad but also infuriating: had IBM not fucked up and not signed Microsoft, that mediocre company that Microsoft is would have gone nowhere and we wouldn't have to suffer their turds since decades.
P.S: in their dying breath (and I know IBM still exists), they did commit one billion dollar to Linux early on (which back then was a big deal). When we see the tens of billions of devices that now run Linux, IBM at least partially participated in Linux taking over the world. So there's that.
PS – Primary sources for the IBM 7950/HARVEST on Bitsavers, excellent as always:
https://www.bitsavers.org/pdf/ibm/7950/