It's refreshing to see a non tech solution every now and then. As technologists we have to learn when and when not to apply technology solutions. This just happens to be one of those cases.
An example he gave was designing a timer for a clothes dryer. The user needs to be able to specify how long to try the clothes. The dryer should run for that long and stop.
The standard solution was a mechanical timer. It had a dial. The consumer turned the dial to the appropriate time, the dryer started, and when the timer ticked down the dryer stopped.
Engineers were starting to replace these with digital solutions. The digital solution was a lot more fun for the engineer. He'd get to design some circuitry, maybe use a microprocessor. He'd need a power supply for the electronics. He could have buttons and a nifty LED display. Maybe have preprogrammed drying times.
That's a hell of a lot more fun for the engineer than to just specify sticking some commodity mechanical timer in there like his father and grandfather would have done.
However, the mechanical timer is cheaper, easier to build into the dryer, more reliable, and easier to use for the consumer. There is no advantage whatsoever to the consumer or the dryer company in using the custom digital solution--it serves no purpose other than the entertain the engineer who designs it.
Analog does not scale with complexity nearly as well as digital.
The question isn't whether or not to use technology. The question is which technology to use.