Curious about human brain development through an animal model? 🧠🐒Check out our new preprint on Brain Charts of the Rhesus Macaque Lifespan! We've mapped normative trajectories for macaque brains using 1.5k MRI scans from PRIME-DE and collaborators! biorxiv.org/content/10.110…

What happens when we break down DSM diagnoses into their individual symptoms and rebuild our classification system from the ground up? Do we see familiar DSM constructs, HiTOP dimensions, or something else entirely? New paper out today! ✨ journals.sagepub.com/doi/10.1177/21… 1/🧵

I'm playing with NotebookLM and feeding it my articles, and it keeps on removing my calls for radical approaches to global language loss, and instead proposes 'raising awareness' as the solution.

The new GSM-Symbolic paper from Apple has been making waves, but we published very similar findings earlier this year. Using nearly the same symbolic template methodology on GSM8k problems, we demonstrated the reasoning limitations of LLMs. arxiv.org/pdf/2401.09395

Elsevier is one of the largest, most hated and most influential academic publishing companies in the world. How it started & how it's going A thread🧵

Connectome: anatomy Network: a bunch of connected neurons circuit: a bunch of connected neurons PCA: beginner level magic Manifold: advanced magic neural dynamics: neural activity dynamical system: anything that changes attractor: stable state in neural activity

Here's the recording of my talk on Language and the Politics of Death from yesterday: youtube.com/watch?v=W4fiy8…

Itai Yanai The paper discusses the traditional format of scientific papers and argues that it misrepresents the actual thought processes involved in scientific discovery. The key findings of the paper are: 1. Observation is never truly "naive" or "unbiased" - it is always shaped by our

We were sad to hear that omnivoreapp is shutting down (though wish the team the best!) We just shipped Omnivore Import so you can get all of your reading data from Omnivore into Reader easily

Any AI technique that is less data-efficient and energy-efficient than human intelligence will inevitably be replaced by something better in due time. Data-intensive and compute-intensive techniques can only be temporary, like the room-size computers of the 1950s.

I don't have too too much to add on top of this earlier post on V3 and I think it applies to R1 too (which is the more recent, thinking equivalent). I will say that Deep Learning has a legendary ravenous appetite for compute, like no other algorithm that has ever been developed