any thoughts on our new review?

Apparently we goofed on PSI - we ought to grovel a bit about that.

Reviewer generally gets it — but makes a huge mistake in thinking Hakaru is just about generative models rather than also inference algorithms. We ought to clarify that quite seriously.

The reviewer sure seems to know a lot about PSI… and have taken quite a lot of time to review this.

we’ve gotten 3 very detailed expert reviews

which i’m quite happy with, even though I disagree with them some

And they came out ABB. And because we’re held to a higher standard, that seems like acceptance is in serious doubt.

The main thrust of the reviews seems generally fine. Too bad the ICFP ratings are so coarse.

I think that’s unclear — the biggest piece of “higher standard” is “have a strong advocate and no detractors”

That we have, as none of the experts are detractors, that can outweigh the lone ‘meh’ report.

I think Ken’s in Boston already (we’ll both be in Salem from late tomorrow until Thursday-ish)

Reviewer E seems to be someone who works on PSI and who doesn’t understand that this paper cannot be tailored for any single person. Probably Timon Gehr. Anyway, I don’t think it’s worth trying to improve this reviewer’s opinion of our paper, such as by reminding this reviewer of contributions they missed. But it may be worth writing a response that helps this reviewer keep their opinion and helps other reviewers improve their opinion. To start, we can answer all explicit questions (to reinforce the impression that we know our stuff), and take this opportunity to reiterate that our system motivates and contextualizes and validates and promotes our contributions but does not exhaust them. I’m not sure what else.