Re: change of pace in a language or conlang
If we take the rules of thumb from glottochronology in the IE languages, languages change just barely noticeably in 100 years, become a new language in a few hundred and are unreconstructable after about 10,000. From watching deaf children spontaneously creating their own languages, those languages evolve rapidly initially and then slow down like ordinary natural languages.
In the corpus, after about 2 years after toki pona's birth, it all starts to look the same and new learners adopt those conventions really fast. So at the moment, imho, toki pona is already at the point where convention and "the mob" is deciding, not so much the conlanger deciding, be it jan Sonja or fans in the community like jan Pije (and the other 2, jan Pije emailed me their names).
I suspect the issue with lojban is not that conlang get to stay in that state of being undecided and flexible for longer-- the case with lojban is that it is so hard that no one can learn the conventions fast enough to get a serious community content to just read and write it, where as toki pona was learnable in 2 months and once one feels competent in the language, any major reworking of it seems like a bad idea or at best, an invitation to repeat two months of work. For the lojban fan-- that probably isn't fluent yet-- it's not to painful to suggest a radical change that would make prior art obsolete.