yes!

it would be interesting if running could support a high quality independent and audience funded outlet like cyclists have in @escapecollective.bsky.social

cite is from a natpost "review".

found this masterclass in misleading citation in the speccie:

so we are somewhere in between embrace and extend but a bit before extinguish

en.m.wikipedia.org/wiki/Embrace...

that's super interesting! is it possible to share a bit more about this, i'd love to use it as an example in my class?

and the new "i'm feeling lucky" standard

Apparently, Russian propaganda directly worked to secure the election of a specific candidate in the U.S. elections in 2024.

it’s like a hack-and-leak but from the inside out.

wow nice photo too!

100%, just like the old google. (leaving to one side llms aren’t info retrieval systems). and the ux will devolve in the same way for the same reasons.

one thing i puzzle over - where will the new sources come from?

"OpenAI says that Deep Research is trained to select solid, reputable sources"

reminiscent of the effort aimed at the syria civil defence, which used youtube/twitter very effectively. much of what we know about how that worked was from the streaming api, which no longer exists. that this is still "dark magic" all these years later is such a failing.

a betrayal of the distinguished tradition of subprime mortgage-backed security safety and radium health & beauty safety!

i think you are right. somewhere in the mix of "what is special about the potential social risk of ai" is the affordances we are giving it as a regulative technology. l'etat c'est moi, translated to machine.

4/ The biggest surprise? Media indoctrination and civil liberties repression are the most predictive of autocratic survival. These findings have big implications for “information autocracies” in the digital age.

reminded me of the “human encyclopedia” from frasier