The way I see it, an engine's an engine. LSx or JDM import...
If I understand you though, it sounds like you're referring to zeroing out an IAT based correction table on a particular EMS or hack of the OEM EMS. If that's what you're talking about, then sure-- zeroing out a correction table that for instance, pulls timing at higher IATs, could in theory boost performance, so long as the fuel is adequate to prevent detonation under the conditions, and since the OEMs generally tune on the conservative side, I'm not surprised at all.
But I think we're talking about something else here.
What the OP asked about if I understood correctly though, was physical sensor placement to make sure it gets an accurate reading of incoming air temps as the air enters the intake. On a Speed Density based EFI system the IAT sensor is a critical measurement. You can zero out the correction table, that's just a correction table though, but try to flatten out or fake out the IAT reading to the ECU and the ECU has lost a huge piece of the equation it uses to calculate the proper base pulsewidth (that that correction table would be correction, if the base PW is wrong it doesn't matter what you do in the corrections, it will never be right). I've done testing on my dyno at different intake temps (easy to do, open and close the hood while drawing air from the engine bay for the engine) with a improperly calibrated IAT sensor and then properly calibrated the IAT sensor to see the difference in real time. I could easily see a 1.5-2 point difference in the AFR when the IAT sensor was miscalibrated (and reporting slightly off-base temps to the ECU, which it then used to calculate fuel pulsewidth incorrectly). If the IAT was gone altogether, then the ECU has no way to properly calculate fuel PW on a speed density system.