The largest collection is in the Bavarian state library, which somehow managed to preserve most of its collection during WW II (although one fifth of its collection, including that of movable type prints, were destroyed). Surprised no one has tried to write a thesis/book about the mechanics.

was that the Vatican should hold the largest collection of incunabula. But turns out, no, it only ranks third, which somewhat surprised me.

This was back in 2015. And if I recall correctly, the main person at the time used to be Misha Khoklov, not sure if he is still around (we lost touch after 2020 or so, which reminds me I should write). Programming used to change every two quarters or so, and there used to be themes.

Yes, I was at the TTI, but also at UChicago. Most of my grad work was at UChicago, and I was fully immersed in the university culture. So a solid 5 years.

We had shown this at Doc Films some years back (I used to program it for a while). If you are in UChicago, get hold of someone at Doc Films, and they can probably get it screened.

AISTATS had the highest quality amongst my stack for the three. Hopefully it is never discovered more broadly to be a (far less popular) moral equivalent of NeurIPS, ICML, ICLR, and like UAI, remains small (and strong).

Bitcoin, also known as leveraged QQQ, is down around 20% in the period after August. Once it starts to bottom out and rise reliably, would be a signal that something has changed in the character of the so-called debasement trade (i.e. either risk-on is back, or it started benefitting from it, or..).

otherwise like to insist (correctly) that stock markets are mostly detached from economic and policy fundamentals (and can be detached for an irrationally long period of time, to paraphrase a dictum due to Keynes). But turns only till the point that a data point buttresses some other larger point.

You mean wireless communication technologies in the category of the bridge? I would say even more so. Expertise there is even more infrastructural and impersonal. The doctor and lawyer case is just that its more individual and personal, and there are stakes.

One might say, well not true, you just drove over a bridge in the morning. That's true, but that gets folded into the background of daily life -- very easy to take for granted, or not even notice.

In both you have a combination of high stakes. v. high informational asymmetry, and a sort of tacit, experience-conditioned judgement that can not be faked (where the outcome is the competence, the consequences of which can be existential). Other domains rarely put one in that position so directly.

A list of all 230, very simply presented. en.wikipedia.org/wiki/List_of...

3:

Round 2:

The mathematical structure here is quite interesting, but probably not worth to get into for a thread. In the meantime, can just use the Conway book as a copout, which covers them in the first ~quarter or so (although it's also a bit too cute).

This isn't exactly trivial given that there are 230 such groups, and you need to be able to search efficiently over them (allowing distortions). As an aside: For 1D they are known as Frieze groups (total 7), wallpaper groups (total 17) in 2D. Same math, just different dimension.