
My article on AI for science, in which I characterize a deviant notion of scientific objectivity rooted in the impossible ideal of theory-free inference, is available now open-access in Erkenntnis link.springer.com/article/10…

I’m not like the other Bayesians. I’m different. Thinks about philosophy of science, AI ethics, machine learning, models, & metascience. postdoc @ Princeton.

My article on AI for science, in which I characterize a deviant notion of scientific objectivity rooted in the impossible ideal of theory-free inference, is available now open-access in Erkenntnis link.springer.com/article/10…

And now, time for my favorite game show! Is it the mercury retrograde or is it the vibecode slopocalypse coming home to roost?

I'm facing the same issue tonight. Every attempt to upload new application materials fails with a "MediaID was unavailable" error, even after trying multiple browsers. With several faculty application deadlines today, this is extremely frustrating. @Interfolio
Has it ever occurred to anyone at @Interfolio that sometimes several institutions have deadlines on the same night and sometimes several people will be submitting materials to those institutions simultaneously. So that it would be a real bummer if the site crashed at that time?

Every first and fifteenth of every month throughout the fall term the “upload PDF” function stops working. A new perk added by @ElsevierConnect for premium subscribers.

Has it ever occurred to anyone at @Interfolio that sometimes several institutions have deadlines on the same night and sometimes several people will be submitting materials to those institutions simultaneously. So that it would be a real bummer if the site crashed at that time?

Spoiler: you can’t “automate peer review” any more than you can “automate democracy.” But the reasons people think you can automate peer review are richly informative about latent epistemologies of metascience and science evaluation!

This Princeton Philosophy panel on the computational universe with Stephen Wolfram should be fun!

Tomorrow I’ll be talking about Automated Social Epistemology: Language Models in Scientific Peer Review with Princeton AI’s Natural and Artificial Minds Initiative

conflicting classes and thrice redefined function calls. your error log is five times longer than the code itself but by some holy miracle nothing has forced it to abort.

it always compiles but is a palimpsest of legacy code copy-pasted and cobbled together like some uncommented frankensteinian golum

You forge a deal with a genie: your published scholarship bestows sanity upon its readership, but whomsoever shall gaze upon your version control and file organization instantly acquires mental illness.

The question on university campuses should now be how to get rid of all the administrators who caved to and collaborated with a grossly unpopular political agenda, including curricular changes, reversals on diversity mandates, buying into AI hype, clamping down on Gaza protests…

importing models from the cognitive and behavioural sciences to their study. Claiming that “AI is like the Shameplant” is simply confusing the de dicto for the de re at a yet higher level of abstraction.

Spoiler, AI (LLMs) is not really at all like the Shameplant (Mimosa pudica). But they are both objects of study that have been ascribed “learning” and “memory” and other cognitive attributes because—I claim—we have failed to exercise the requisite degree of caution while

Rahul and I talk about model transfer, conceptual hygiene, and why AI is like the Shameplant.
Surrendering Science to the Machine? Discussion with philosopher of AI and machine learning @bayesianboy.
