GT sapphire glaziers: You signed WHAT deal with Apple?

Dear lord, who let these chumps out of the playpen

Naturally occurring efficiency

At that point, all was settled and done, or at that's what we thought. Except for two little problems. The first of which was that GTAT doesn't actually seem to have been competent at making the sapphire itself. Which is, if you're setting out to make sapphire, something of a problem.

For previous to this GTAT had been a manufacturer of machines to make sapphire. It made the furnaces that made the boules, and it made the cutting machines that cut them. And it obviously had some experience of how to use them: you can't design and build machinery without at least some clue as to how to use it. However, clearly there's a little bit more to it than that, and we shouldn't be all that surprised about it actually. An awful lot of knowledge is local, some of it is implicit, and a great deal of both are gained simply by doing.

There's one puzzling economic finding out there, that factories just seem to naturally get more efficient over time. Even with the same (or one that changes gradually over time) workforce, the same machines, the same organisation on paper, making the same things in apparently the same way, factories can become 20-50 per cent more efficient over a couple of decades of operation.

Sapphire scratchproof glass is a crucial component of all-manner of Apple products

Sure, we also get great leaps in efficiency when someone thinks up an entirely new way of achieving the same task, but it is a bit of a puzzle as to why it just seems to happen, almost all on its lonesome.

The explanation usually given for this is just that the organisation of the place, the “culture” if you like, begins to reflect that implicit knowledge of what it's learnt by doing. Might be something as silly as everyone takes their teabreak at 09.30 instead of 10.00. Once you've loaded up the machines, got them running, checked everything, there's a natural break at that point and so why not stop then and come back at 10.00 when all the machines need adjusting, instead of standing around for 30 minutes and then having a break just when all the machines do need adjusting.

Or maybe it's the butcher's van for the canteen which works out that a delivery at 07.30 makes more sense, as he gets away quicker and isn't fighting the first wave of morning delivery trucks. You can make up your own examples but we've all had to show the new gal in the office the way that the server works faster if you kick it just there, just then. Nobody knows why but everyone does know it.

That's the first problem we can identify: that GTAT just didn't appear to have any of this learning-by-doing knowledge regarding how to actually run banks of the machines that it constructed. It may have been pretty shit hot at machine making but not at machine running.