@neil @NAB @proactiveservices One of the bugs listed in the Horizon technical appendix is phantom transactions tracked down to touchscreen; so that's probably the closest to actual direct hardware related failures in this case; there are other things (like crashes and timeouts) caused by dodgy hardware or comms, but the failure involved there is when the recovery software fails. I don't think there's any evidence here of any other 'hardware gives wrong result' errors.
@penguin42 @neil @NAB @proactiveservices well apart from Intel's CPU that couldn't divide correctly, for example
@falken @neil @NAB @proactiveservices Hey I used to have one of those (I got it replaced under Intel's scheme) - long before this Horizon stuff; but also it was a very obscure error in pretty small values as I remember. You'd have some fun proving that any of today's intermittent CPU errata make it through to a user.
@falken @penguin42 @neil @proactiveservices
But doing what it was programmed to do by a human?