niplav @niplav@schelling.pt

TED talk:

"Imagine a printer..."

"that works."

crowd goes fucking insane

first you learn quicksort and mergesort, then you go back and learn insertion sort and selection sort. then, you learn heap sort.

and you never learn bubble sort.

this is the "machete order" for sorting algorithms.

Keith is an incurable romantic

“If you really cared about this AI thing, you’d replicate Ted Kaczynski’s successful strategy.”

what if we held hands 👉👈 🥺
at the Sloan Great Wall
and listened to the pulsars sing

I notice I can just *not engage* with the Discourse and do my job

blahaj is mid

you are trying to solve the right problem :) with the right methods :) based on a correct model of the world :) derived from accurate thinking :)

Hot take: Making poorly calibrated public predictions is pretty awesome.

Not only can you learn from them, but it's fantastic that you were willing to predict something concrete clearly enough to actually be wrong - which is unfortunately rare.