Freek Oude Maatman
@freekoudemaatman.bsky.social
Theoretical(ly) Psychologist, Philpsy/Philsci |Joint PhD student @University of Groningen @Radboud University | Theory evaluation | Psychopathology | Complexity | he/him
Should they take place within the Utnapishtim itself, or should they be in-universe/independent from the Utnapishtim?
June 11, 2025 at 9:03 AM
Should they take place within the Utnapishtim itself, or should they be in-universe/independent from the Utnapishtim?
Congrats Mel!
June 6, 2025 at 6:27 AM
Congrats Mel!
Sorry, I hoped a parody might lift your spirits! 🥲
And it is depressing indeed: I would have joked I can't wait for fully automated research pipelines (from RQ generation to publication) to cut out people using LLMs, but I heard yesterday those are actually in development already. 🙃
And it is depressing indeed: I would have joked I can't wait for fully automated research pipelines (from RQ generation to publication) to cut out people using LLMs, but I heard yesterday those are actually in development already. 🙃
May 14, 2025 at 3:05 PM
Sorry, I hoped a parody might lift your spirits! 🥲
And it is depressing indeed: I would have joked I can't wait for fully automated research pipelines (from RQ generation to publication) to cut out people using LLMs, but I heard yesterday those are actually in development already. 🙃
And it is depressing indeed: I would have joked I can't wait for fully automated research pipelines (from RQ generation to publication) to cut out people using LLMs, but I heard yesterday those are actually in development already. 🙃
"🚨 New in metascience! 🚨
AI-generated preregistration reproducibility recently was proven to be low, due to LLM instabilities.
As a solution, I and my colleagues propose meta-prompt preregistration:
pregistration of prompts used to generate preregistrations, specifying LLM version numbers..."
AI-generated preregistration reproducibility recently was proven to be low, due to LLM instabilities.
As a solution, I and my colleagues propose meta-prompt preregistration:
pregistration of prompts used to generate preregistrations, specifying LLM version numbers..."
May 14, 2025 at 2:43 PM
"🚨 New in metascience! 🚨
AI-generated preregistration reproducibility recently was proven to be low, due to LLM instabilities.
As a solution, I and my colleagues propose meta-prompt preregistration:
pregistration of prompts used to generate preregistrations, specifying LLM version numbers..."
AI-generated preregistration reproducibility recently was proven to be low, due to LLM instabilities.
As a solution, I and my colleagues propose meta-prompt preregistration:
pregistration of prompts used to generate preregistrations, specifying LLM version numbers..."
Pfft, sorry for (possibly) suggesting something you already know! 😅
In my experience, everything DAG-related flows down to @dagophile.bsky.social at some point (perhaps through several bathtubs?).
In my experience, everything DAG-related flows down to @dagophile.bsky.social at some point (perhaps through several bathtubs?).
April 9, 2025 at 8:02 AM
Pfft, sorry for (possibly) suggesting something you already know! 😅
In my experience, everything DAG-related flows down to @dagophile.bsky.social at some point (perhaps through several bathtubs?).
In my experience, everything DAG-related flows down to @dagophile.bsky.social at some point (perhaps through several bathtubs?).
I may be wrong, but his sounds like the Iwasaki & Simon (1994) solution! If so, and if you are interested in a less mathematical, DAG-based version, this treatment by Naftali Weinberger is great!:
philarchive.org/rec/WEINAT-8
philarchive.org/rec/WEINAT-8
Naftali Weinberger, Near-Decomposability and the Timescale Relativity of Causal Representations - PhilArchive
A common strategy for simplifying complex systems involves partitioning them into subsystems whose behaviors are roughly independent of one another at shorter timescales. Dynamic causal models clarify...
philarchive.org
April 9, 2025 at 7:58 AM
I may be wrong, but his sounds like the Iwasaki & Simon (1994) solution! If so, and if you are interested in a less mathematical, DAG-based version, this treatment by Naftali Weinberger is great!:
philarchive.org/rec/WEINAT-8
philarchive.org/rec/WEINAT-8