3 ms·
My gut tells me they may never end up being very useful. something about the intrinsic downsides to any analog computer. which if you are not familiar, is that
by CamperBob2 3y ago
My gut tells me they may never end up being very useful. something about the intrinsic downsides to any analog computer. which if you are not familiar, is that analog computer components must be far more precise than their digital counterparts.
Counterpoint: some of the most interesting classical algorithms, such as transformer-based neural nets, appear to work well with far less precision than expected. For that matter, the first perceptrons were analog... and of course, so is the one that everybody is trying to emulate.
As a measure of signal-to-noise ratio, numerical precision in the classical domain may correspond to other attributes in other domains. It may not be a productive thing to focus on.