
I think this part of the argument isn't true. If natural selection is a blunt tool, gradient descent is a science fictional molecular scalpel by comparison. There's extremely high bandwidth between the objective and the changes to the model, which natural selection does not have.

Liron Shapira @liron

The inflation adjusted version of the employees per $1M chart implies that worker productivity plateaued right around when smartphones were invented.


🧵 People in the Past were Desperately Poor 🧵
1/
Whenever I discuss history its shocking how much ignorance I encounter from even trained historians about the level of poverty in the past
Contrary to idiots and revisionists No. People in the past were horrifyingly impoverished


for the longest time the top google search completion for "why won't" was "why won't my parakeet eat my diarrhea," because any human being who saw that phrase would immediately click, and not for the reason we wanted them to

elena 🌸 @cheer_wine

Fascinatingly, AIs bullshit (speak without regard to truth) under the same conditions as humans.
Folks bullshit when they are expected to have an opinion on a topic they don't know & they won’t be corrected.
AIs hallucinate most when forced to answer on topics they don't "know"



First time in my life i've ever heard of Oakland police catching a hit-and-run driver. usually bad drivers in Oakland are arrested when they leave the city and get caught by like Alameda or Berkeley PD.


“In 1700, almost 1 in 25 inhabitants on Earth, and one in five in Europe, was French. Today, less than a percent of humanity is French”
If you want to understand why, and the long-run consequences of this, read my piece in online magazine @WorksInProgMag


It's been very encouraging to see @NSF's willingness to experiment with novel funding mechanisms.
Letting reviewers occasionally take "non-consensus bets" on projects they are especially excited about could accelerate scientific breakthroughs.
