Neither is this. An airplane measures indicated airspeed by sampling dynamic pressure. This happens to be useful because it has a direct correlation to various

Neither is this. An airplane measures indicated airspeed by sampling dynamic pressure. This happens to be useful because it has a direct correlation to various important V-speeds, such as the speed where a wing in a certain configuration generates enough lift for forward flight and so on. This is why an aircraft's V-speeds are given in KIAS. (Put another way, you get the density altitude calculation for free by relying on IAS, as the airplane gets more "inaccurate" with altitude but in actuality "accurate" from the perspective of fluids traveling over surfaces.) As an airplane gains altitude, IAS dramatically falls versus true airspeed due to this measurement artifact. Complicating this calculation for someone approaching it physically, as you have: by regulation all aircraft calibrate to the same altimeter pressure (29.92 inHg) beyond 18,000 feet. If you think about this physically, that means a whole bunch of airplanes indicate 35,000 feet but in actuality linger around 34,000 or 36,000 or something depending on the actual atmospheric pressure. That also affects how airspeed is calculated, meaning you can't say "well, at 33,000 feet, Mach 0.9 is blah." That computation is a little more nuanced based on the way aircraft sensors are set up and, until computers, required a whiz wheel. 'mikeash is correct, here. You're correct for TAS. IAS is a weird aerospace-specific thing where it's actually helpful to be wrong. I know. To the point of the article, too, the computer worked to avoid exceeding Vne and probably started intervening after passing Vno, which are both figures given in KIAS because both rely on how thick the air is. (Everything in this comment applies to subsonic flight, and the entire calculus changes when you're >=Mach 1.)