Getting the location + gradients of divergences (not of iteration starting points)


Re: core team commitment to output redesign: idk who exactly is the “core team” but Bob and Mitzi both already stated this is a priority (for various reasons unrelated to this thread), and Dan and I have both tried to push things along in this direction at various times (including now) with moderate success so the time it’s right w.r.t. developer interest. I’m not completely sure which design we’ll end up with but that’s just how open source goes.

… wasn’t Michael (?) the one who wrote code (mcmc_writer or similar) to bundle output in the mcmc code to begin?