I'm not suggesting to include new strains on the fly while they are discovered without trials. The same way they had to do trials for Gardasil-9 after already having Gardasil, they could do the same just with a wider cocktail.
In the optimal case you would attempt the 100+ cocktail and don't have any unreasonable side effects from them. Of course if there is a single variant in there that causes side effects you would need a ton of trials to triangulate which one causes them, but there is no guarantee that that is necessary or that you wouldn't run into that variant when you just adding 5 new variants in an iteration.
Even if he is wrong about how it works, because of the potential for unknowns they need to study it in a trial. There is a lot we don't know about biology.
This is both true for natural infection and first generation vacines.
Later generation vaccines avoid this by not presenting certain proteins.
My point is, you can't just include hundreds of protein variants without a lot of trials.