But we don't have the details, do we? Suppose cryonics preserves the graph of neural connections; we can look back in 10000 years and know exactly how many neurons you had, and exactly which were connected to which others. Is that enough information to revive you, or someone who's 90% identical to you? Who knows? We really have no idea.

Your point about redundancy is, I think, looking at it from the wrong angle. I would expect brain redundancy to handle random errors like "lost 2% of neurons", but the idea that we would have multiple fundamentally different mechanisms for encoding memories seems evolutionarily implausible. If we simply haven't preserved anything about, say, the thicknesses of the glial cells, or we know which synapses are present but it turns out they have different sized gaps and that's important and we can't recover that information, or any one of thousands of other pieces of brain biology that might turn out to be vital for encoding memory, then cryonics won't work.