we all know I have no life, so:

I think it was actually with the 2009 mixes

also why there are no stories where a man has to care for an unrelated child

she’s right, this is why no one has ever had a paternal figure in their life who wasn’t their biological father

I'm as sciencey as they come for philosophers. But I got into philosophy because, in some important sense, the man on the Clapham omnibus is right. You are allowed to look at a scientist, say "what the fuck are you on about?", and they (or at least us philosophers) have to say something sensible

all I want are Beatles remasters that update the whack stereo mixing but leave everything else unchanged

it’s genuinely bizarre that republican hawks aren’t clamoring for hegseth’s resignation

like it sucks to override independent districting commissions but hopefully that will make them more attractive to both sides in the long run

yeah I was pretty iffy on the Dem redistributing thing as a matter of principle until I realized this point. only way you get mutual disarmament is if R’s realize there’s a credible threat to them

whenever people are like “this official economic statistic doesn’t capture people’s real economic situation” it’s like, ok, then why did it *use to* predict economic sentiment, and what changed?

this is the thing that always drives me up the wall when people find some new statistic that they say captures people’s “real” economic situation. ok, so does that statistic predict *past* economic sentiment?

in general I think it is often reasonable to be less deferential to a given discipline’s methodological commitments

though as an outside observer it does sometimes look like certain possibilities get excluded bc of unmotivated methodological commitments

I also enjoy when they make causal claims but treat counterfactuals as poison

yeah I think in general you can trust their first-order findings

probably one of my most chauvinistic takes re: philosophy’s relation to other disciplines is that a lot of historians *really* need to study more epistemology

my edgy Covid take is that I while I did socially isolate until vaccines came online for the good of society, I absolutely loathed being unable to see people in person, throw a party, or travel for over a year, and I enjoyed absolutely nothing whatsoever about any of it

one thing that sucks about the current moment is that every time harms are mitigated there’s a part of me that’s like “shit, now no one’s gonna learn their lesson” (and that includes Dems)

Voting is a collective action problem. To make it more likely that people will vote, you need to reduce the net cost of voting. A candidate you have to hold your nose to vote for raises the cost. A candidate people are excited about lowers the cost of voting.

what do you think the relationship is between Hume’s problem of induction and Goodman’s riddle of induction? same problem, different problems, one includes the other, etc.?

There's nothing inherently Jewish about any of this. Put any group of people in the role of master over slaves, and they'll start to act like it.

that's what the pretest is supposed to control for though

we have the comparison in performance between the control condition and the AI condition, and we have the comparison in performance between subpopulations within the AI condition. I don’t see how that tells us how the subpopulations in the AI condition would have performed in the control condition

wait how?

but then the “didn’t use” group significantly outperforms *the control*

wait the results aren’t making any sense to me. among people in the AI condition the “direct solution”, “hints”, and “didn’t use” groups all performed similarly, which according to the authors indicates similar intrinsic skill & motivation

hmm I have to think about this

wrt this result, the problem is that “participants who don’t just ask AI for the answer” is a self-selecting group, so that introduces all sorts of confounds, e.g. intrinsic motivation to do the task

participants in the no-AI condition can remember what was in the reference panel after it was taken away. also I’d need to pull up the paper again but it doesn’t look like they measured whether no-AI participants *used* the reference panel, which would be the relevant comparison

though see some additional context — still don’t really think the control condition is comparable though