Define "massively wrong". My personal opinions (stated w/o motivation for brevity):

• Building AGI from scratch is likely to be unfeasible (although we don't know nearly enough to discard the risk altogether)
• Mind uploading is feasible (and morally desirable) but will trigger intelligence growth of marginal speed rather than a "foom"
• "Correct" morality is low Kolmogorov complexity and conforms with radical forms of transhumanism

Infeasibility of "classical" AGI and feasibility of mind uploading should be scientifically provable.

So: My position is very different from MIRI's. Nevertheless I think LessWrong is very interesting and useful (in particular I'm all for promoting rationality) and MIRI is doing very interesting and useful research. Does it count as "massively wrong"?