@cogscikid.bsky.social
210 followers 74 following 59 posts
Kempner Institute research fellow @Harvard interested in scaling up (deep) reinforcement learning theories of human cognition prev: deepmind, umich, msr https://cogscikid.com/
Posts Media Videos Starter Packs
Pinned
Excited to share recent work on an AI model that doubles as a scalable cognitive theory.

Basic idea: when learning one task, "preplay" other tasks to learn implicit goal-oriented maps that support fast performance later.

This improves AI and predicts human behavior in naturalistic experiments!
Reposted
Every class on machine learning should study the datasets upon which the field was built. They provide more insights than any probability bound.
Benchmarking our benchmarks
The only validated theories of generalization are sociological and historical.
www.argmin.net
Reposted
every single tech idea is like “soon our robots will be capable of playing catch with your kid, freeing you up to spend more time working on your employers’ spreadsheets”
Reposted
I'm recruiting PhD students to join my new lab in Fall 2026! The Shared Minds Lab at @usc.edu will combine deep learning and ecological human neuroscience to better understand how we communicate our thoughts from one brain to another.
Reposted
Come work with us! UC Irvine Cognitive Sciences is looking for a new Assistant Professor to join our team: recruit.ap.uci.edu/JPF09896

I'm not on the committee, but happy to talk if you're interested.
Assistant Professor - Cognitive Sciences
University of California, Irvine is hiring. Apply now!
recruit.ap.uci.edu
Reposted
@arthurpr4t.bsky.social has written another tour de force, showing how efficient coding reshapes numerosity representations in parietal cortex.
Reposted
We have got exciting (and unconventional) stuff cooking and we are hiring for a strong research engineer on the GDM Game Theory team in London.

Consider apply if you are interested in the intersection of game theory, multiagent systems and LLMs!
job-boards.greenhouse.io/deepmind/job...
Research Engineer, Game Theory & Multi-Agent Systems
London, UK
job-boards.greenhouse.io
Reposted
🚨NEW: For the first time in history, the single largest group of people arrested in the interior and sent to ICE detention is people with NO criminal record.

As of yesterday's data release, 40%(!) of those arrested in the interior and detained by ICE had no prior conviction or pending charges.
Chart shows Detention Population, Among Those Arrested by ICE in the Interior, by Criminal Record. The chart shows three lines; Prior conviction, pending criminal charges, and no criminal record. From 2019 through September 2025, the largest group is people with criminal convictions, with much further down people with pending charges. But starting in January, those without criminal records spikes dramatically. In July it crossed pending charges and in the latest data it jumps above criminal convictions.
Reposted
Brown’s Department of Cognitive & Psychological Sciences is hiring a tenure-track Assistant Professor, working in the area of AI and the Mind (start July 1, 2026). Apply by Nov 8, 2025 👉 apply.interfolio.com/173939

#AI #CognitiveScience #AcademicJobs #BrownUniversity
Apply - Interfolio {{$ctrl.$state.data.pageTitle}} - Apply - Interfolio
apply.interfolio.com
the truth is we have a terrible education system so we leave people unequipped with the ability to think critically about these things

the USA has made this mess with how it treats public education

the good news: better public education is bipartisan

the bad news: its not a priority for elites
Reposted
Excited to share a new preprint based on my work this past year:

**TreeIRL** is a novel planner that combines classical search with learning-based methods to achieve state-of-the-art performance in simulation and in **real-world autonomous driving**! 🚘 🤖 🚀
Reposted
How can scientists apply efficient #AI tools to cognitive science? And what does it mean for how robots learn in new environments? #KempnerInstitute fellow Wilka Carvalho (@cogscikid.bsky.social l) + collaborators outline a new framework in 3 papers.
Read more: bit.ly/47Fk9F9
#NeuroAI
Excited to share recent work on an AI model that doubles as a scalable cognitive theory.

Basic idea: when learning one task, "preplay" other tasks to learn implicit goal-oriented maps that support fast performance later.

This improves AI and predicts human behavior in naturalistic experiments!
1/2 the problem is demand-driven labor/"jobs"

the other half is picking skills to grow in

UBI (as your article points out) can give the breathing room for growth

but the space of skills to learn (and the economic "demands" for lab) will be vast and opaque
Love the question.

"jobs" where people contribute to society are one way people commonly get "purpose"

I assume that a world with a dramatically different labor market will be one where people will need to reimagine how they contribute to society
fair. that's an important part of the social welfare solution.
bearish on UBI unless it's coupled with a tool to help people find activities that give them meaning/purpose
So when we need to incorporate this knowledge into decision-making, there really is no replacing humans
Even in conversation, we constantly use knowledge we can't explain. Sensing emotional undercurrents in someone's voice. Knowing when to pause. Feeling when someone needs encouragement versus honesty.

Right now, LLMs have little to no access to this knowledge
Same with dancing - the rhythm in your body, reading your partner's movements. Or coffee tasting - not just "bitter" or "smooth," but the complex flavors that build a sommelier's expertise.
How do you teach ice skating through text? "Bend your knees, shift forward, arms out for balance." But what about the feel of the blade catching ice? The micro-adjustments your body makes? The sensation of momentum?
A limitation of LLMs: tacit knowledge---the knowledge you can't put into words

LLMs have all the world's written knowledge. But many important things were never written down.