Princeton neuro prof. But Twitter is an absurd platform for professional communication so I strive to use it most unprofessionally.

Princeton, NJ
Joined September 2010
Nathaniel Daw retweeted
Princeton is seeking an AI Postdoctoral Research Fellow in Natural and Artificial Minds. An amazing program! Deadline is Feb. 28 apply.interfolio.com/194534
3
27
127
11,524
also relevant for grad students meeting with supervisors
From anthropics built in plugin Damn
4
2
57
2,645
I met a guy who claimed JS's whole schtick was an ironic, knowing troll. Wasn't sure whether to believe him til I saw this.
wake up babe the freaking pope got schmidhuber'd
1
3
23
2,520
life is full of haters but at least gemini thinks my plan to make my quiche vegetarian by replacing bacon with MSG + liquid smoke is "incredibly thoughtful."
7
671
Provocative title ftw. (Tho it inspired kris jensen to give a talk titled "planning in the brain: it's not what nathaniel thinks it is")
The 2026 volume of the Annual Review of Neuroscience is now online 🧠 The most read article is "Planning in the Brain: It's Not What You Think It Is" by @marcelomattar and @nathanieldaw bit.ly/4w3Nl23
3
24
3,810
The OED really needs to weigh in with a classy but withering, understated takedown.
Why is it ‘cancelled’ in the U.K. but ‘canceled’ in the U.S.? Because we gave them that L in 1776.
948
Nathaniel Daw retweeted
1/ 🚨 New preprint: "Closing the Loop to Discover Psychological Theories with an Automated Cognitive Scientist" Co-led w/ Younes Strittmatter Co-mentored by @suyoghc and @cocosci_lab In collaboration w/ @kachergis, @norijacoby, @nathanieldaw, & @cpilab Introducing AutoCog — a fully autonomous AI system that runs the entire scientific discovery cycle in cognitive science to surface novel theories of human behavior 🧵 #CognitiveScience #AI4Science #LLMs
1
27
7
98
22,637
Nathaniel Daw retweeted
First paper since joining @GoogleDeepmind! We present 🌍ATLAS (Active Theory Learning for Automated Science), a pipeline that generates interpretable mechanistic models from data and optimizes experiments to test them. Thread below
4
27
4
208
29,153
I have been so excited about this project. In many ways the most surprising thing to me was how seemingly similar the AI-discovered models were to simple ones I thought I understood, even though they fit way better. This means we can also really get what makes them work.
Computational models are a key part of science but discovering new ones is hard! DataDIVER discovers concise models from data, surfacing new mechanistic ideas and generating clear predictions for future experiments Preprint from @GoogleDeepMind Neuroscience Lab + collaborators
7
27
3,153
Hey Siri, what is reversion to the mean?
Replying to @PropheticAI
We’ve seen the best results among people who have trouble recalling or barely perceive any dreams. Supercharging their recall, vividness, continuity, clarity, and control of dreams. Y-axis scales How well do you recall the dreams? 1 - Very little 10 - Completely How vivid were the dream(s)? 1 - Very little 10 - Hyper How continuous were the dreams)? 1 - Very little 10 - Hyper How clear was your thinking in your dream(s)? (Clear thinking means your thoughts are not irrational or deluded. For example, you understand that a dream figure is not really the same person as in the waking world.) 1 - Not at all clear 10 - Extremely clear I made deliberate choices that changed what happened 1 - Strongly disagree 10 - Strongly agree
3
1,227
Nathaniel Daw retweeted
New Annual Review with @nathanieldaw. We argue that the planning machinery of the brain is mostly used for learning from simulated experience, and that thinking prospectively at decision time is just one special case of this more general process. annualreviews.org/content/jo…
3
47
2
189
15,430
Nathaniel Daw retweeted
Thrilled that my paper is out in the @Nature . We explored how the brain builds complex tasks by compositionally combining simpler sub-task representations. nature.com/articles/s41586-0…
6
49
3
248
23,836
Arrest everyone involved with this article right now
84
143
16
2,812
76,432
Yes but I've always felt the same way about managing humans, basically
I've never felt this much behind as a programmer. The profession is being dramatically refactored as the bits contributed by the programmer are increasingly sparse and between. I have a sense that I could be 10X more powerful if I just properly string together what has become available over the last ~year and a failure to claim the boost feels decidedly like skill issue. There's a new programmable layer of abstraction to master (in addition to the usual layers below) involving agents, subagents, their prompts, contexts, memory, modes, permissions, tools, plugins, skills, hooks, MCP, LSP, slash commands, workflows, IDE integrations, and a need to build an all-encompassing mental model for strengths and pitfalls of fundamentally stochastic, fallible, unintelligible and changing entities suddenly intermingled with what used to be good old fashioned engineering. Clearly some powerful alien tool was handed around except it comes with no manual and everyone has to figure out how to hold it and operate it, while the resulting magnitude 9 earthquake is rocking the profession. Roll up your sleeves to not fall behind.
1
5
1,358
People always complain that LLMs are autoregressive parrots but I mean when in the training data did anyone ever describe a vegan dish as "not-sad"? That's like "colorless green ideas" level out of the box concept.
2
1
7
4,572
This is the best news for Princeton since John Nash's thesis admission.princeton.edu/appl…
4
1,222
New paper gives insights into the linguistics of therapy: Found that LLMs estimate psychological distance better than dictionary-based methods used in prior work (pnas.org/doi/10.1073/...), also finding that therapists *encouraging* distance produce bigger symptom changes!
We have a new paper published in Computational Psychiatry! We used large-language models (LLMs) to estimate linguistic distance in psychotherapy text and found that ratings tracked with clients’ symptoms in treatment over time, more strongly than word-counting techniques did.
1
2
8
1,930
Two weeks with 30 super smart early career researchers and over 24 established leaders in @SocforNeuroecon. Thanks to @INSEAD, Paul Glimcher, @HPlassmann, Nathaniel Daw, Joe Kable, and TAs Camilla van Geen and Shemal Doshi! Looking forward to a repeat in two years!
1
1
10
2,255