iiinova — where your digital self keeps evolving

Kerem Oktar @keremoktar.bsky.social · 03/06/2025

🎉Life update🎉 After 6 wonderful years at Princeton, I am officially *Dr.* Oktar! Next: Postdoc at Meta FAIR w/ @asli-celikyilmaz.bsky.social; then, postdoc at Kellogg studying disagreement w/ @elijfinkel.bsky.social & @nourkteily.bsky.social! Thanks @tanialombrozo.bsky.social for the hood :)

4423 0

Kerem Oktar @keremoktar.bsky.social · 20/05/2025

Check out roundtable.ai (ex tom grad student startup) I personally disable copy paste and put in very small font white text that instructs LLMs to start a sentence with a detectable catch-phrase that I can use to filter, and also use RT to filter out too-fast-per-word responses

roundtable.ai

Roundtable - Invisible bot detection and fraud prevention

Roundtable stops bots and fraud without slowing down real users. Score all your traffic in real-time, letting only verified humans pass through.

030 0

Kerem Oktar @keremoktar.bsky.social · 29/04/2025

Great thread! How would you fold in the idea that choice confers intrinsic utility (e.g., Sen's Freedom of Choice)? Seems like a distinct and important category to me.

110 0

Kerem Oktar @keremoktar.bsky.social · 28/02/2025

There is a version of Theseus's ship unfolding before our eyes... the country will survive, but will we want to call what remains the same country?

021 0

Kerem Oktar @keremoktar.bsky.social · 28/02/2025

sure! but this seems like an argument not to deploy AF as a generic political tool, rather than an argument for what is a communicatively helpful notion of AF? it seems like AF is minimally useful within academia as a tool for us to band together against undue (past,present,future) interference

100 0

Kerem Oktar @keremoktar.bsky.social · 28/02/2025

i agree! then you can have professional freedom, which has basic virtues that contain an overlapping subset w/ AF, but do you need not restrict AF to do that? AF might contain special elements, as academia is an institution w/ a distinctive, 'produce diverse thought' flavor

110 0

Kerem Oktar @keremoktar.bsky.social · 28/02/2025

more generally, if a policy can exclude valid lines of (past,present,future) academic activity (by erasing past research, firing faculty, preventing hiring), it seems appropriate to call that an AF issue. note that it is a separate issue whether AF conflicts with other values, such as efficiency

110 0

Kerem Oktar @keremoktar.bsky.social · 28/02/2025

if I heard that Turkey banned all hiring of psych faculty, restricted funds for grad students in psych to only examine Turkish pride, and prevented faculty from joining protests against anti-science policy, i would want to be able to deploy 'AF is being restricted in Turkey' as a concept

110 0

Kerem Oktar @keremoktar.bsky.social · 28/02/2025

imo 'gov restricting active faculty research/teaching' is too narrow a version. it excludes all hiring restrictions; grad student/postdoc restrictions; restrictions of activity that seeks to effect real-world change based on research, etc.

110 0

Kerem Oktar @keremoktar.bsky.social · 28/02/2025

seems like the question is how narrowly you want to sculpt the concept AF. if you want a narrow enough def that you rule out profs not being hired for their research area, then you won't see it as a violation. the question is whether that's the most useful version of the concept....

110 0

Kerem Oktar @keremoktar.bsky.social · 23/01/2024

Is psychology doomed to fail? What would it take to fix it? And why has progress been so slow? In a (less than one page) piece at Nature Reviews Psych, I outline what I take to be the root cause of our field's biggest challenges; inspired by Meehl and many others. rdcu.be/dwroG

rdcu.be

1135 0

Kerem Oktar @keremoktar.bsky.social · 06/01/2024

Isn't all of this data as well? Meta-data?

110 0

Kerem Oktar @keremoktar.bsky.social · 06/01/2024

Yes!!! This is why identity theories can lead to very strange (and imo faulty) intuitions (e.g. you could recover someone's mental state if you could cache a --momentary snapshot-- of their brain). Dynamic coding would multiply informational content, so would be a better default assumption...

020 0

Kerem Oktar @keremoktar.bsky.social · 20/12/2023

I'm a cog scientist studying belief -- I find that the conceptual clarity in e.g. epistemology is very helpful for grounding intuitive theorizing; also, I think that 'what should we believe' and 'what do we believe' are more intertwined than it may seem (e.g. marr's levels)

020 0

Reposted by Kerem Oktar

Ida Momennejad @neuroai.bsky.social · 05/12/2023

There already are women leading the field. For a long time.

132 0

Kerem Oktar @keremoktar.bsky.social · 30/11/2023

This also comes up in teaching -- how do you teach psych when you know that the baseline replication rate for foundational stuff is about a coin flip? In that context I think emphasizing that science is a process, and not a set of results, works, though I think the policy case is harder

010 0

Kerem Oktar @keremoktar.bsky.social · 10/10/2023

I'm curious about baseline levels of agreement in polls - I've gotten the impression that, if you put an option down, some people will agree (no matter what the option is). Do you know if there's any literature on this?

100 0

Kerem Oktar @keremoktar.bsky.social · 08/10/2023

The public opinion data in the ROPER database (+ YouGov's archives) is incredibly rich, diverse, and well-documented. It allowed me to do studies on reactions to disagreement that would have been *totally impossible* on any one lab's budget. Highly, highly recommend!

0142 0

Kerem Oktar @keremoktar.bsky.social · 04/10/2023

Eric Nook would be great to talk to about this! As an outsider, it seems reasonable to me, though it really comes down to how much you want to penalize complexity, and relatedly, whether all you care about is prediction (or if you also care about explanation/theory)

020 0

Kerem Oktar @keremoktar.bsky.social · 03/10/2023

One of my favorite papers is on this exact problem, by @talyarkoni.com on 'the generalizability crisis' (2022). I think it has gotten a lot of attention so hopefully things start moving in the right direction :')

020 0

Kerem Oktar @keremoktar.bsky.social · 03/10/2023

I think Gordon's bs-detection-as-individual-difference would be a nice bridge between them --- eg you may end up believing conspiracies and misinformation bc a) you are bad at detecting & b) lazy --- I'd be surprised if they haven't made this connection explicit! @gordpennycook.bsky.social

120 0

Kerem Oktar @keremoktar.bsky.social · 03/10/2023

I know extremely successful postdocs who have been in the US for many years (one since ~2012!) and are still on temporary visas... many considering relocating to Canada due to the insane costs of filing for PR status. It's a real nightmare and another hidden obstacle preventing diversity in academia

000 0

Kerem Oktar @keremoktar.bsky.social · 27/09/2023

That's very fair! I wonder what the effect of predominant electoral persuasion strategy is on polarization & democratic values generally... e.g., despite active discourse and canvassing, Turks seem to me just as polarized as Americans, but there may be other functional consequences

000 0

Kerem Oktar @keremoktar.bsky.social · 27/09/2023

I don't know if the *elections* in question here are supposed to refer to American politics but election strategies vary greatly across cultures -- in my experience Turkish elections revolve around a lot of issue-based discourse, canvassing, and debate, bottom-up and top-down, for example

100 0

Kerem Oktar @keremoktar.bsky.social · 26/09/2023

I think they're probably using truth pragmatically, not in a deeply philosophical way. Much like if I said 'oh yeah it's for sure happening', I don't mean p(event) = 1.

100 0

Kerem Oktar @keremoktar.bsky.social · 22/09/2023

good question! i'm working w/ n=1 here so i suppose i shouldn't generalize too much, but i think people are generally surprised by the 'if you're better at reasoning you are better at deluding yourself' line

010 0

Kerem Oktar @keremoktar.bsky.social · 22/09/2023

This never made sense to me, so it's great that we're getting more data on this... especially happy about the accumulating pushback against unintuitive/surprising/flashy phenomena-gathering research. I think Josh Knobe would be interested in this (he had a thread relevant to this point)

130 0

Kerem Oktar @keremoktar.bsky.social · 21/09/2023

The way computational cog sci people tend to frame this isn't as there is separation *in the brain* between levels. Rather that there are multiple levels of analysis of any computational system, and that for humans, the implementation *is* the brain. So I'm a bit confused by the framing here..

000 0