It may actually be faster since GDDR6X has complicated timings and fake bandwidth that relies on a bunch of voltage levels to transmit data and I don't believe it does that as efficiently all the time..
I don't think that a different encoding/signaling scheme is actually "fake bandwidth" - DDR5 does something similar and it does tend to deliver, but a ~5% reduction in memory bandwidth can be compounded by a ~5% increase in default clock speeds or by the activation of an extra SM partition. With 20 Gbps G6 vs. 21 Gbps G6X, if Nvidia opted for an eventual extra SM route (bringing it up to 6144 cores) - I think I would prefer the G6 version in this situation.
Should be called as 4060 Ultra or 4070 SE. Now this reminds me of the infamous GT 1030 DDR4.
The G5 and D4 versions of the 1030 were so radically different in performance that you could call the D4 a 1010 (which eventually released in the OEM market) or even a 1005 in comparison, but in this case it's practically a nothingburger. If they compensate it somehow by increasing engine clock or activating an extra core partition, it's gonna balance itself out and the G6 version may even enjoy reduced power consumption. In any case, I don't think it'll be a showstopper.
You were driving with his wife too?
Whoa, I'm not that kind of guy
