This actually makes a hell of a lot of sense. The 'deal' with diff gears, in practically any production diff, is that the pinion isn't in the centre axis of the crownwheel, it's down below it. That means the gears don't mesh like a gearbox gearset does, they actually slide across one another. This generates big loads, and potentially, lots of heat (and it's also why you need to run hypoid spec oil - hypoid being the eng term for the offset pinion/crownwheel gear arrangement - and it's also incidentally why you _don't_ want to use hypoid oil in a conventional gearbox, as some of the stuff required to make hypoid gears survive isn't as ideal for conventional gears). If you are curious about the heat aspect (and the heat the gear surfaces run compared to the oil can be a lot hotter) - remember more than a few race cars run pumps to route fluid to and from the diff to a cooler. Another interesting tid bit, diff oil viscosity is rated on a different scale/system than engine oil, and when it's up at operating temps, 'thick' diff oil pours/flows similar to something like castrol gtx2 engine oil!

On the 'straight cut' option - believe it or not helical gears in a gearbox (if they are made of the same stuff and are the same basic dimensions, and similarly finished/treated) are actually stronger than straight cuts, they just 'lose' a little bit of power as some of it is wasted due to the helix angle trying to force the gear forward or backward in the gearbox casing, ght cuts transmit it all to the next gear (or as much as is practical). On a limited (relatively) power NA racer, you need every % you can find, hence their use. they are also frequently made to a much higher standard than mass production stock engined cars gearsets would be!

In terms of percentage benefit of polished gears, I doubt you could find a massive difference in power output (you might if you kept it on the dyno long enough to really start cooking the diff oil, which would be 'great' for the rest of the engine/driveline!!) but it'd certainly be (at least potentially, and of course it's down to how bad they were to start with, how well it was done, how close to failure the std sets were etc) possible to extend longevity as much as 2-3 times on a circuit racer differential's life. Mostly because it'd be preventing the oil overheating toward the end of a race, and once that happens, even small amount of use, like 1 lap, causes massively accelerated wear.

These comments are general from stuff I've had any exposure to, not necessarily datto specific, and I'll happily stand corrected if there was other stuff going on, I don't mean to speak for the original poster in any way!