Sönke.ehret
@s_ehret
Metrics and data. I do computational social science. Work on applied cultural evolution, social norms, strat comms. Views are my own.
579 Following 211 Followers 493 Posts
Latest posts 5
Analyst results as a distribution. Finally!
it's seems obvious to me that we're going to do an automated review/replication of *all* of social/behavioral science, both empirical & theoretical. We'll bypass papers where it's not possible, but it will be noted & an asterisk will be applied. The Great Replication...View quoted post on X
lol
Benoit Mandelbrot, the father of fractal geometry:
a 10-sigma market crash has odds of roughly 1 in Avogadro's number. by the Gaussian math taught in every business school, it simply can't happen.
"It should never happen. It happens every day, somewhere, to somebody."
October '87? a 23-sigma day. and 10-sigma moves, he says, are "all over the place."
his verdict on the Gaussian risk model taught in every business school: "sheer idiocy."
~70-min lecture, free. the mathematician Taleb calls his master, on why markets break the math ↓View quoted post on X
Interessanter Punkt. KI-Regulierung hat viel mit Fantasie und Vorstellungskraft über die Zukunft zu tun. Und diese kann man gut beeinflussen zb in dem man auf extreme Edgecases hinspielt.
republik.ch/2026/06/05/eur…
republik.ch/2026/06/05/eur…
www.republik.ch «Europa soll aufhören, das Silicon Valley zu kopieren!» Ein Interview mit der Tech-Journalistin Karen Hao über KI-Modelle der Zukunft.