Colder engine (ie 160-180 deg thermostat) makes more horsepower? Maybe he meant to say colder air into the engine makes more HP.
+3 mpg for a K&N conical air filter? Looks like they moved the air intake from a cooler source to a hotter one....smart. A 4" 90-degree air inlet elbow is really restrictive? (I'd bet the air filter is much more restrictive). A wider snorkel inlet to that 4" air inlet makes more airflow and hp? (I suppose if you put a 2" flute on your garden hose it will flow more water). (Hey, why not route that air inlet down to a cooler place and ram it in from the grille/wheel well?...that will make horsepower will real/cool rammed air).
That guy's engine bay is no prize. After listening to that, I'm not sure there was a single piece of accurate information given other than stating the engine is a "giant air pump." Both of my cars have aluminum heads. In fact the last one I owned went 230K miles w/o warping.
You are absolutely correct regarding the location of the Air Inlet to a warmer area being a mistake. It is not easy (without careful measurement at all operating conditions) to determine what inlet temperatures underhood actually are; for example they may change at speed. Since there is little need for power at slow speeds it may be OK to be at a seemingly hot area that might be cool at speed, for example.
Also it is very dependent on the pressure available at the location, which would require testing with a manometer. You won't know the answer without testing. It is very vehicle-specific as well; there is no "one size fits all". The most obvious example is the "Cowl Induction" location at the base of the windshield; in some cars this is a low pressure area (bad), in some it's high (good).
I do disagree about the filter being more restrictive than the path, however. Bends in the path are quite restrictive while the typical OEM air filter flows more than the engine actually needs in stock form. On the other hand most OEM intake paths now incorporate Helmholtz Resonation optimization ... in layman's terms this is an acoustic chamber that acts in conjunction with the engine timing (ECM, plus Cam timing) to improve torque at some specific RPM. Removing that feature can reduce torque output from the engine. Again, very vehicle specific. (A Helmholtz Resonator in the intake looks outwardly somewhat like a plastic catalytic-converter shaped section).
The ideal filter location and path should be at the highest pressure area, typically in most cars this is at the front licence plate location. Depending on the shape of the vehicle body there may be other areas, but testing is required to determine where they are. The more aerodynamic the vehicle's shape, the less likely an alternate location can be found, and most modern cars have quite low drag coefficients, reducing the likelihood you will find an alternate.
Getting air from there to the motor's intake in a straight path or one with minimum bend severity is far from a trivial matter (the car's radiator tends to be right in the way, and it's hot itself), while any benefit requires at least freeway speeds, if not higher.
Because the stock filter is typically more than "good enough" the aftermarket filers like the K&N often offer no improvement at all. This can be improved, but not without careful intake tuning, and not, typically, on engines without power improvements over stock form. Put another way, any potential improvement in airflow can't be achieved without corresponding engine modification.
A (slightly) colder thermostat can make more power, but once more, not on a stock engine. Changing to a cooler thermostat is likely to put the ECM into an un-optimum tune in stock form; they expect to see the 195 thermostat and don't work well without it. A cooler thermostat in a race-prepared engine requires an aftermarket ECM (ideally) or modifications to the OEM ECM (band-aid method) in order to take advantage.