Gotta disagree there... I think Tesla logs whether it's steering, brake, or stick that disengaged. I'd REALLY HOPE that the steering correction is logged and reported as part of snapshots and disengagements (w/ the resulting snapshots). So, I disengage with stick when I mean "I want to go somewhere you don't" (know it's going to take the wrong lane, take a bad route and want it to reroute, etc). I disengage with wheel to say "Yo the lane is over here", and the brake to say "Yo, how about not sniffing that car's butthole".
Most of the time, I'm not surprised by TACC still running, and intend for it to do so. Sometimes, though rarely... TACC is on and it accelerates unexpectedly, and I brake that too. I respond really quickly to AP being stupid, not like most videos I see where the driver is ... pretty much letting FSD do its thing...
Hard to tell what the correct behavior would be, but honestly I'd lean towards steering breakout = TACC disengage as well, being something Tesla should implement. If braking disengages steering (and often, I wish it wouldn't), then having steering not complement disengaging acceleration as well is asymmetrical. Both modes ought to complement each other.
tl;dr: Tesla most likely uses the mode of disengagement as data, but I really believe they should disable TACC as well when disengaging by steering wheel.
I got the impression from watching Chuck Cook videos that using the stalk to disengage was a way to disengage without logging it. Essentially when you want to stop using FSD but don't want Tesla to view it as a disengagement, use the stalk. I personally have not heard from Tesla that this is the case; I am assuming that as an early-early FSD tester, he was given more info on how disengagement works.
Agree with the annoyance of TACC staying on after wheel disengagement. A few of us have asked Tesla to address that.