niplav @niplav@schelling.pt

Forecasting is Worse is Better

The Brier score was invented in 1950, Bayes' theorem was first described in 1763, and probability theory was properly invented at the beginning of the 18th century.

However, attempting to use forecasting to seriously determine the probability of future events, training & selecting people for their forecasting ability and creating procedures that incorporate both intuition-based and model-based predictions have only been around since the mid 1980s

@Paradox just hanging out and working on stuff, getting to know the locals

Im in Berlin now until next Wednesday

I'm now also back on twitter, handle niplav_site (celebrating the fact I managed to use the domain I now use: http://niplav.site

Degrowth leads to stagnation not collapse I think

Site down till tomorrow bc DNS propagation

You gotta pump this agent's money
That's some rookie trading

Sir cum flex

You can't do an 'r' with a circumflex, my horrible buddhist joke falls through

<convex hull of all possible takes on trans people>

Mid people trying to look really hot in photos makes them just look bad

"a little defection never hurt anyone" he said, more to himself, sprinting for the next yad stop

@PraxisOfEvil i do not believe one second that tech billionaires aren't screwing around like crazy

@PraxisOfEvil just a front for the harem of 35 teenagers they're screwing

Classification of simple finite pre-agents

{fuzzy, util} × {terminal, instrumental}

Should i try to meet gwern in person?

@aran follows gwern to learn about AI