Mark Chonofsky
@markchonofsky.com
cyclist, dog enjoyer. currently @wbez.org / @chicago.suntimes.com
loving proprietor of @ctaalerts.bsky.social
loving proprietor of @ctaalerts.bsky.social
Not sure I understand this one, but the argument about 'not X - but Y' being essentially an artifact of greedy sentence construction makes a lot of sense to me
November 27, 2025 at 9:24 PM
Not sure I understand this one, but the argument about 'not X - but Y' being essentially an artifact of greedy sentence construction makes a lot of sense to me
Oh interesting, thank you!
November 27, 2025 at 8:33 PM
Oh interesting, thank you!
Is KL Kullback-Leibler? Do models optimize directly for KL divergence?
November 27, 2025 at 7:49 PM
Is KL Kullback-Leibler? Do models optimize directly for KL divergence?
Sorry, but isn’t it the point of that scene that the extra mark doesn’t do anything? That they’ve just divided the same scale into more segments, like cutting a pizza into more slices - you don’t get more pizza
November 21, 2025 at 2:47 AM
Sorry, but isn’t it the point of that scene that the extra mark doesn’t do anything? That they’ve just divided the same scale into more segments, like cutting a pizza into more slices - you don’t get more pizza
I think LA’s looks worse - en.wikipedia.org/wiki/Metropo.... There are people there serving short sentences, as in Brooklyn, but it’s mostly a jail
Metropolitan Detention Center, Los Angeles - Wikipedia
en.wikipedia.org
November 16, 2025 at 12:07 AM
I think LA’s looks worse - en.wikipedia.org/wiki/Metropo.... There are people there serving short sentences, as in Brooklyn, but it’s mostly a jail
I misread the excerpt - whoops! The 40% was a win number, not an estimate, so there is no causal claim in the cited statistics.
November 4, 2025 at 9:10 PM
I misread the excerpt - whoops! The 40% was a win number, not an estimate, so there is no causal claim in the cited statistics.
This might be because (1) their canvassing worked, (2) the 40% estimate was wrong in the aggregate (high salience election + turnout is hard to model), or (3) of their sample, the people they managed to contact were unusually likely to vote
2/2
2/2
November 4, 2025 at 9:08 PM
This might be because (1) their canvassing worked, (2) the 40% estimate was wrong in the aggregate (high salience election + turnout is hard to model), or (3) of their sample, the people they managed to contact were unusually likely to vote
2/2
2/2
As someone who used to work in this field (on the other side) what they probably mean is that of the several million people they intended to reach, they turned out 50% of the people with whom they spoke. 1/2
November 4, 2025 at 9:06 PM
As someone who used to work in this field (on the other side) what they probably mean is that of the several million people they intended to reach, they turned out 50% of the people with whom they spoke. 1/2
yes and workforce upskelling
November 1, 2025 at 11:52 AM
yes and workforce upskelling
Reposted by Mark Chonofsky
apparently Knol’s founder was inspired to launch an ambitious encyclopedialike entity bc he felt online sources , such as Wikipedia, lacked info about —of all things — New Yorker cartoonists. then the New Yorker licensed cartoons for knol articles web.archive.org/web/20080730...
October 29, 2025 at 2:37 PM
apparently Knol’s founder was inspired to launch an ambitious encyclopedialike entity bc he felt online sources , such as Wikipedia, lacked info about —of all things — New Yorker cartoonists. then the New Yorker licensed cartoons for knol articles web.archive.org/web/20080730...
My understanding is that it's a state constitutional issue, requiring two years to amend, so we wouldn't have new maps until 2028
October 26, 2025 at 5:18 PM
My understanding is that it's a state constitutional issue, requiring two years to amend, so we wouldn't have new maps until 2028
I could do a whole thread on my favorite Muppets moments. I come back to the Jim Henson Tribute again and again (www.youtube.com/watch?v=SRSp...) "I feel very sorry that your best friend Jim died. -- Wha? He died? Jim DIED? But we were just starting to get to know him!" Staggering humanity.
Just One Person - The Muppets Tribute to Jim Henson
YouTube video by DidymusBrush
www.youtube.com
October 22, 2025 at 6:55 PM
I could do a whole thread on my favorite Muppets moments. I come back to the Jim Henson Tribute again and again (www.youtube.com/watch?v=SRSp...) "I feel very sorry that your best friend Jim died. -- Wha? He died? Jim DIED? But we were just starting to get to know him!" Staggering humanity.
We’re not asking for any identifying information from people submitting info
September 13, 2025 at 1:16 AM
We’re not asking for any identifying information from people submitting info
or as I like to call it, Mister Rogers’ Neighborhood
September 10, 2025 at 4:52 AM
or as I like to call it, Mister Rogers’ Neighborhood
they called me and they said it's imperative that you figure it out
September 1, 2025 at 5:27 PM
they called me and they said it's imperative that you figure it out