Out of all the reviewers, Gordon is one of the only ones worth anything. Constructive, intelligent, and not a sensationalist. Adult enough to be nuanced and careful even when criticism is warranted, and able to realize when it is not.
His conclusions make a ton of physical sense. The amount of capability in this camera generates real, problematic heat, just like any decent test shows where people aren't forcing it or disabling safeties. Canon is using 3 sensors to protect internal components for long life of the components, just like any other modern high speed circuit that throttles or shuts down (hard drives, CPU, tons of things have thermal limits). Overheating ASIC and sensors repeatedly can and will degrade them.
Very sensible that ambient temperature is measured in order to predict record time as a function of starting temp plus expected temp rise over time in that mode. That isn't actually a bad way to do it if you want to keep a conservative safe limit for the dies of the chips and sensor. Sure, if the product didn't have the huge sensor and processor activity and strong IBIS and weather sealing, it would cool better and run longer or indefinitely, but we are back to square one where any intelligent testing shows real thermal issues and the product does what Canon says without any conspiracy. People can hack all they want or complain that Canon should sample the temperature constantly and raise the limit, but without the whole engineering picture it is irresponsible to say we know better.
In my own thermal tests in hot sun, the camera was quite predictable and did overheat, and disabling IBIS and the EVF made things cool much faster and last longer. It makes perfect sense to me that even idle, just running the processor and EVF and IBIS, produces enough heat inside the sealed body that must flow away from the sensitive components where temp sensors will be placed. So no conspiracy after all, just a sub optimal algorithm that has now been improved. Can it be improved more without compromising safety and long term reliability of the camera? Can lighter codecs help? How many people actually need this and cannot just use an external recorder vs all the whining and negativity?
Sadly the comments on Gordon's video are the same confirmation bias trash we knew they would be be already: proof of the cripple hammer, fake timer, blah blah blah. I'd say these people are all paid shills for another brand, but the truth is sadder than that- average video crowd is just that dense and spoiled.