It’s Coherence Time
When I was in school, I used to fantasize about building the best gaming machine on the market. At that time, a Pentium III based machine was top of the line and an ATI or Geforce video card would allow you to play Half Life without stutter. There was always something new on the horizon and I’d regular visit the site Tom’s Hardware to read up on the hardware rumor mill.
For a long time, processor speed was the metric to follow. 400 MHz? Meh. 1 GHz — wow! That breakthrough came via the AMD Athlon around the year 2000. A few years after that milestone, it became clear that processor speed was only one element towards a snappier computer. It wasn’t until a decade later that I first experienced running a laptop on an SSD and I recall thinking how much my obsession over the processor speed was insignificant compared to hardware read/write speed.
Nowadays, it’s hard to know how much of an improvement processor speed actually has on day-to-day computing. High end CPUs top out at 5.5 GHz but they can have a dozen or more cores running in parallel. It’s hard to appreciate the effect that this actually has on how things run.
Along comes the potential breakthrough of quantum computing and most people, to some extent myself included, don’t fully appreciate what’s involved in a quantum computer or what breakthroughs it will lead to. I went through the effort of reading over Azure’s Quantum service and Amazon’s Braket and am not really sure what I can do yet with them.
What’s also difficult to understand is how much progress is actually being made in what’s being offered. Is progress made in the number of qubits that can be “solved” at a given time? The physical entanglement distance? Maximum coherence time? Error rate?
Maybe we’re really beyond the “feeds and speeds” era? A quantum computing’s output might be measured in its breakthroughs vs. any given technical spec. One computer might be a drug discovery device (protein folding). Another computer might be a bitcoin miner (if this is the case, please name it “uh oh!”). Still another computer might be the next evolution of natural language and image generation AI.
The problem of communicating progress of quantum computing is that the communications follows a binary 0 and 1 state. The state 0 is where it is now. Lots of companies doing interesting things and promising the moon. The state 1 will be when we’re actually on the moon. The companies try to market themselves as though they’re both on the path to get there and there already.
That message isn’t very coherent.
