People did a test yesterday in IRC: downloading the exact same data for the NeuralNetwork, yet the resulting neural hash differs. With each new user it can make it difficult for consensus (e.g. "because there was 4 new CPIDS in the space of a couple mins").
When the NN wallet ends up with a high rate of same hash it gets popular and a wallet then can create the superblock.
Making it also more difficult is stuff like these (WCG outage. having e.g. 25 superblocks (for each whitelisted project) would prevent that when a BOINC project gets fucked, it "influences" all others by delaying the SB...
I have been observing since yesterday an increase in netweight (currently it is at 91 million), probably it's an increase since like 17h... so, if that doesn't help soon, I wouldn't expect a superblock to come that fast.
How to help?
- come to irc and talk with the guys who investigate this (ravon, BullShark, Gunde, ...), check out also comments in bug 392.
Thanks for the in depth explanation. Makes a lot of sense. If there is anything else we can do as the common folk, sing out and I'm sure we'll do our best.