In 1994, a professor in Virginia found a bug in the Intel Pentium chip that caused one in 9 billion floating point computations to be in error. Everyone agreed that very few people would ever see the error, but Intel handled it the way Apple is handling the (apparently) far more serious problem with the iPhone, resulting in a massive recall that cost Intel $475 million. There was wide consensus at the time that it was a PR bug more than a technical one. If they had not resisted, had been more empathetic, the cost would not have been nearly so great.
Intel was surprised to learn that users expected their computers to be computationally perfect. Engineers at Intel must have known, previously, that they weren't. That's what they were trying to "explain" but the users didn't want to hear it.