The VAX in My Pocket
My first employer ran Florida's telecom on millions of dollars of mainframes and VAX clusters. My phone outclasses all of it. We never stopped to notice.
By Geordie Everitt
This morning I made a video call to a stranger on another continent. It cost nothing. Our faces and voices traveled at the speed of light, and neither of us thought about it for a second, because nobody thinks about it anymore. Later I watched an episode of Star Trek filmed in 1966, remastered so cleanly it looks like it was shot last year on modern equipment by someone going for a deliberately retro aesthetic. The original audience saw it through static on a 19-inch tube. I saw it sharper than its own editors ever did.
I want to put some numbers on how strange this is, because I happen to have lived on both sides of the transition.
Eight Gigabytes for a Hundred People
My first full-time employer was GTE Data Services — GTEDS — now long since folded into Verizon. It processed telecommunications for much of Florida. It ran several IBM 370 mainframes and operated the largest VAX cluster in the southeast United States. This was serious infrastructure: raised floors, halon fire suppression, an operations staff that worked in shifts around the clock.
I worked in Technical Publications. The department had about a hundred employees, a full print shop, and a hundred VAXstation 4000 workstations connected over 10baseT ethernet, reaching the mainframe through a PDP-11 that served as a gateway — one obsolete computer employed full-time as a translator between two others.
The entire department shared eight gigabytes of storage.
Eight. For a hundred people producing documentation continuously, with revision histories and print masters. We managed it the way you manage any scarce resource — carefully, politically, with meetings about it.
The flat I'm writing this from, in Spain, has more storage and more bandwidth than that entire company. The phone in my pocket outclasses the mainframes, the VAX cluster, and the PDP-11 put together, and it does so while also being a camera, a television studio, and a device for arguing with strangers.
The LaserDisc Future
Here's my favorite part, because it shows how the future actually looked from inside the past.
GTE had spent eighty years and billions of dollars building out telephone land lines. Naturally, the company wanted to find new things to sell over them. One experiment I remember was video-on-demand: movies stored on LaserDiscs, played back across the telephone network to your home. Racks of LaserDisc players, robotically swapping discs, pushing analog video down copper pairs.
It seems quaint now. It wasn't. It was a rational bet made by intelligent people extrapolating from the infrastructure they had. The mistake wasn't the engineering. The mistake was assuming the scarce thing would stay scarce — that bandwidth and storage were permanent constraints you designed businesses around, rather than temporary conditions about to evaporate.
Today there exist videos, many hours long, consisting entirely of puppies playing in front of fake jazz, and people stream them in high definition while asleep. Megabits per second, gobbled by the unconscious. The bandwidth that GTE's engineers rationed like wartime sugar is now spent on content explicitly designed not to be watched.
I can think of no cleaner measure of abundance than that.
Nobody Registered the Transition
The strange thing is not that this happened. Technology improves; everyone knows the Moore's Law catechism. The strange thing is that it happened without anyone really pausing to register it — there was no moment when we collectively looked up and said: the constraint that organized the entire industry, that justified the buildings and the shift work and the meetings about disk quotas, is gone.
Each step was incremental. Each year's capacity was only somewhat more than the last. And so a transformation that would have read as science fiction to my colleagues at GTEDS — your phone will outclass the data center, and you will use the surplus to watch sleeping puppies — arrived without ceremony, the way a tide comes in while you're reading on the beach.
This matters for a reason beyond nostalgia. We are standing in the early years of an identical transition — this time not in storage and bandwidth, but in something harder to name. The mainframes of cognition are being miniaturized. And if the last transition is any guide, we will fail to register this one too, until someone writes the equivalent of this post in 2060 and marvels that we ever held meetings about it.
GTEDS needed thousands of people to keep its machine running — programmers, operators, and an entire department of writers documenting it all. What happened to the work when the machines stopped needing it is a different question from what happened to the machines.
That's the next post.