Assorted Links

Always start with politics, except when a squirrel happens.

The OpenAI project started with a big boom, but whether this is a good idea we don’t know yet. People sponsoring the project know about the orthogonality of intelligence and values. Why then an open-source AI project is the best way forward, I don’t really know.

It seems that economists are people and do not answer coherently independent of the wording of questions.

Scott Sumner wrote The Midas Paradox and somebody explains why the book is important. It seems to be a good way to be a politician as an economist. The book is generally well received in the econblogosphere (is this still a word?).

I must reshare this link: The most terrifying statistics question. Every programmer loves their sorting algorithms, but not all sorting is made equal.

Better to learn about some probability distributions to embiggen your talking. Thanks HN.

Burritos are not the best explanation for monads, and other great insights: Start with examples, move to the abstract, and hurt your brain in the process. No way around it.

This week in stupid tech. Sometimes it’s (hopefully only) gigantically bad luck (HN), sometimes it’s bad business (HN, HN #2), sometimes it’s bad committee (don’t forget about the recent ZigBee HA screwup…).

Back to good tech. Or at least trying-to-be-better.

Bayesian inference is all we need, so let’s create a magic bayesian inference DB (HN).

Rayon is a data-parallelism library with fast fork-join primitives for Rust (reddit/rust), that while still being unoptimized seems to be good enough for first uses.

Haskell happened to evolve in 2015 (reddit, HN).

And last but not least, PureScript intros are there to be found. I didn’t know about this language yet, but it seems to be a well thought-through approach for strong functional programming with JavaScript, and later WebAssembly as back-end. PureScript extends on some of Haskell’s concepts, and provides the Eff monad as an improvement to everything is IO, and also provides some more finegrained typeclasses (HN). Any similarity to the like of the interface segregation principle are surely purely coincidental. However, the typeclass universe is far from closed yet, although until recently I didn'even know that in 2015 we don’t have an accepted solution to provide the numerical tower in our programming languages.