With apologies if I'm repeating what was already said.
I've seen this claim on DPReview, usually some Canon user bashing Pentax' SR, or some clueless "internet expert" proving that he doesn't know what a SLR is. What I didn't see is any sort of proof.
First, the only way for SR to impact precise focusing and exposure, given the SLR design, is by using processing power. Since both the PDAF system and the metering system are independent, and fixed in relation to the camera body - there is otherwise no difference if the camera has a SR system or not.
I'd say the (negative) impact is negligible, if any. After all, we've seen improvements in AF speed as Pentax started using the SR.
About the OIS helping significantly with the AF, if it was true we would have proof - Canon and Nikon users observing a decrease in AF speed just by turning the IS/VR off. Same lens, same camera: it would be trivial to test. We have no proof of a positive impact, at least in normal situations.
Note: I'm talking specifically about the AF speed, not about easier hand holding. And of course the example with the plane is about hand holding. In very low light, close to the working limit, I have to keep the camera very stable so it will lock on the target instead of hunting; this is an edge case where OIS might help. This is still about hand holding.
By the way, I briefly tried OIS with Pentax, and my K-5 didn't focus like a D4
As for metering: I don't buy it. Given the working range, and how the metering system has a much coarser resolution than the imager, I don't see why we would get gross errors attributable to camera shake.
Long exposures are a red herring in this case; the exposure system works down to -3EV regardless of which ISO (and corresponding aperture and shutter speed) you're using.
No matter what you do, you cannot convince the metering system to take seconds for a reading.
The actual proof of an impact would be in weird inaccuracies happening only when hand holding, but not with the camera on the tripod. I never noticed the metering going crazy in low light.
---------- Post added 12-10-17 at 02:16 AM ----------
As for "Why VR is built into the lens", the answer is obvious: because Nikon (and Canon) introduced stabilization before digital, with their film SLRs. There simply wasn't any alternative to OIS: moving the film to correct camera shake was not a practical option.
When digital came, their choice was already made.
Last edited by Kunzite; 10-11-2017 at 04:17 PM.