When the recommendations are in conflict with the guidelines

From stunning interview with michael yeadon by el gato malo.  The subheading is uncommonly clear eyed discussion of why the covid vaccine development strategies don't make any sense.

he raises 3 main points:

to create a vaccine with ongoing efficacy, you should pick the part of the virus that is most genetically stable so it is least likely to mutate away from induced immunity. this means you would never pick the spike protein. (though, as he admits, perhaps this was not known early on, but i also suspect there were strong intuitions in that regard given the reams of data from gain of function and other coronavirus manipulations/serial passaging etc)

you would also pick the part of the virus that is most different from humans. this is a key to side effect mitigation. if you train immune response to something that looks like parts of you, you get auto-immune response. your body attacks itself. this would also lead to not picking spike protein which has some similarity to many parts of humans.

in vaccine design, you select for things that are not toxic in their own right. this, again, points away from using spike protein.

and so we have 3 points that, to use michael’s words “teach away” from selecting spike protein as the target for vaccine design.

and yet, pfizer, moderna, astro zeneca, and JnJ all picked it.

why?