In the end it's a win-win for consumers. GTX 1060 can be all this better than RX480, but if they don't price it somewhat close to RX480, they won't sell any because most people aren't willing to pay that much.
Considering the price difference of previous models, I would guess GTX 1060 will be priced around $250 or less. When it gives better performance, way better efficiency, probably better OC headroom, etc. This will probably sell more than RX 480, 470 and 460 combined... (Remember how good GTX 970 and 960 sold?)
The bigger problem for AMD is that GP106 will be smaller and cheaper to make than Polaris 10. It's also more power efficient, so Nvidia can create a wider selection of binnings out of the chip. So AMD can't push Nvidia on price without loosing more money.
We wouldn't have seen this card, especially this early, without the RX 480.
GP106 was taped out late last summer, so this is exactly right on track. It was supposed to be released before "back to school shopping", so this is exactly as expected.
Competition is good, but this is
not an example of an effect from competition. The launch of new chips are pretty much
predetermined nearly one year ahead, it can be postponed but it can't be pushed forward, that's technically impossible. Several chips of GP106 has been in QA for a couple of months*, which is the strongest sign of a pending release, so
we knew the launch was imminent. The same goes for the hype about Vega, some fans are pretending it will be released 6 months ahead of schedule, which simply can't happen...
*) BTW, we also know a GP102-400 is in QA, so get ready for a new Titan soon.