On Fri, Jan 8, 2016 at 4:22 AM, Bryan Ford <brynosaurus@gmail.com> wrote:

Andy Isaacson wrote:

...

The privaTegrity (PT) backdoor is significantly more malignant than the Tor dirauth issue.

[Massive snip in from worthwhile tor-talk thread on central points of authority / DHT...] Are there networks where there is an autonomous AI engine within the network itself, a possible distributed execution sandbox, that decides what, likely changing over time, set of nodes will serve elevated roles? And where AI's programmed evaluation metrics may somehow be tied to some kind of consensus wherein like blockchain, major portions of net must agree or risk fork. And how self-adaptive and permanent is the AI?