Ben Recht
@beenwrekt
Blog: Webpage: The Irrational Decision: How We Gave Computers the Power to Choose for Us. (Princeton)
What do you mean it's not straightforward to show? This all seems obvious to me, bro....
It could only happen on Twitter. You really need militant rationalist brain and powerful cognitive dissonance to defend constantly screening yourself for cancer.
In 2023, I removed the RDD line fits. This result has been suspicious from the get-go.
Here's the key figures from the talk. The first shows the estimated effect of the shingles vaccine on shingles. The second the estimated effect on dementia.
God, that blog post is extreme dogshit. Also... say what you will about pangram but
Staring into the ten-year chasm between the Sam Frank and Sam Kriss features in Harper's...
I can see the beneficence, but I'm having a hard time squaring this with respect for persons. (Amazing how frequently sentences like this appear in methods sections of psychology papers.)
The number of people who actually read the paper that Claude generated, divided by the number who breathlessly liked and retweeted... He posted page one. Check out page 6.
Very curious about the behind-the-curtain ops used to handle these events.
So they made a list like this, invented a bunch of heuristics, and thought they were solving intelligence. But they didn't realize the limits of their own methods. 2/2
Over on the Gelman blog, there's a post about this histogram of z-scores in published papers. I'm always baffled by the discourse around it. Why is this plot a symptom of a problem? Under what model would we expect this to be a bell curve?
The list of signatories on the latest "Please Skynet, don't kill us" letter is BONKERS.
Hah, that's a fun takeaway. Let's also not forget that the author of that substack is a horrible dork too. At the same time Bubeck was lying about GPT websearch, that guy was bragging about his nonsense AGI psychometrics paper with hallucinated references.