1)K10 has more changes than core2 duo had over it's predecessor.
2)Looking at the modifications it's a huge improvement over K8.
Examples are:
128-bit wide SSE units (double that of K8)
512-entry indirect branch predictor and a larger return stack (size doubled from K8) and branch target buffer
More aggressive instruction prefetching, 32 bytes instruction prefetch as opposed to 16 bytes in K8
Wider L1 data cache interface allowing for two 128-bit loads per cycle (as opposed to two 64-bit loads per cycle with K8)
And many more. It's logical to assume it will be more than a little bit better than k8.
3)K8 isn't as terrible as you seem to believe.
4)Penryn is nothing dramatically better over conroe. Just a bit of extra FBS and clockspeed. You're better off taking a E6600 and overclocking it.
5)Codename K10.5 (I think it's basically like what intel did for quadcores but for 8 cores)
Here's
289 pages of K10 optimization to read through (I won't pretend I read it all)