taion

One distinction is that Goldman's SecDb graph is really optimized around offline derivative pricing and risk calculations, which has always been one of the core strengths of Goldman's quant platform. There is a slight distinction between reactive graphs used for those approaches v. those that are more targeted toward real-time online applications.

The comparisons to Galton's studies were dominated by selection biases. They're not indicative of much of anything in particular, except perhaps how not to do this kind of research.

http://slatestarcodex.com/2013/05/22/the-wisdom-of-the-ancie...

There's an extremely extremely rough utilitarian analysis that says that the largest positive effect in QALYs added comes from fewer road traffic deaths due to people substituting away from driving while drunk to driving while stoned. See section IV of http://slatestarcodex.com/2014/01/05/marijuana-much-more-tha...

This really isn't practical in New York City. Almost all of these spaces are leased, even the ones for institutions like Union Square Cafe. There are very few restaurants here where the owners own the building as well.

I agree that it's extremely depressing. I think we actually need to go farther than unbundling on the last mile, though. Ideally, where possible, we ought to go farther than just re-using existing copper links and instead set things up so that we can more easily and cheaply deliver FTTH.

This net neutrality victory is something to celebrate, but we need to keep the bigger goals in mind for this to not be suboptimal longer term.

Isn't last mile unbundling a bigger deal than Title II for ultimately delivering better broadband service? Common carrier provisions are obviously very good, but excluding the possibility of last mile unbundling seems like a bad thing for the longer term.

Are you doing anything clever to figure out e.g. which subset of changefeeds subscribed to a table might be interested in a given update?

Let's say I have a table containing data for many users, while each subscription only needs data for a single user. Instead of scanning through all the changefeeds, you could put subscriptions in a hashmap and figure out which ones to update in O(1) time rather than O(N) time in number ob subscriptions, per update.

Obviously this is much harder in the general case, but do you do anything along these lines?

You can get a very basic real-time push in PostgreSQL with per-row triggers and NOTIFY/LISTEN (pg_notify/pg_listen).

The hard part is building out a change feed that lets you synchronize on a subset of a table in a way that is performant, and to make this work with joins (i.e. when you're using the database as more than just a document store).

Ah, no wonder, thanks! It still seems like quite the coincidence that so many groups were independently (?) working on describing images with CNNs hooked up to RNNs, though. I guess it's just an idea whose time has come?

Why did all of these papers come out today? Did some sort of deadline/embargo lifting pass today?

It's also a bit weird because NVIDIA are usually pretty generous in giving out free K40s to academics. Wonder why they couldn't get some from NVIDIA.

It's not a data store, but doesn't Meteor's functionality set get you pretty close to the same thing? You can essentially build the same type of behaviors on top of Meteor as you can with Firebase.