This strikes me as partially implausible. When the liquidity rewards were disabled, witness pay increased from 25% to 33% (rough numbers) as a share of rewards, but its amount did not change. It was not 10% previously, and I find it difficult to believe you intended this given the obvious numbers of 1 STEEM each for content, curation, witness and liquidity in the original design (later some of curation was shifted to content). That this is 25% each, and not 10%, is blatantly obvious.
What was and is 10% in the white paper (and I assume the code?) is the long term portion of rewards allocated to witnesses (after the minimums are reached) at 0.75% of market cap per year. Long term, this seems fine, if possibly a bit low, but I see no merit in slashing funding currently going to witnesses considering that most if not all of this funding is currently being used to fund valuable work that is helping to develop and fill gaps in the platform, both of which are sorely needed at this early stage. I do not agree that posting rewards, being designed as a lottery and shared with social media users, are generally suitable for funding platform initiatives. What is your proposed substitute for funding the valuable work that witness pay currently supports?
The myth that @recursive and a few others have been disseminating about witnesses being primarily embezzlement and fictitious jobs is just plain false trolling, and you should should not be influenced by it (and if it were true, 10% would be too much)
I support decaying/expiring votes and increased backup witness pay. The latter has been a concern of mine all along (given the expected increases in operating costs I did not see how backups could be viable at all with such low pay), and I'm happy to see it addressed.