Konrad Kording
@kordinglab.bsky.social
@Penn Prof, deep learning, brains, #causality, rigor, http://neuromatch.io, Transdisciplinary optimist, Dad, Loves outdoors, 🦖 , c4r.io
It may take a bit
November 14, 2025 at 8:08 PM
It may take a bit
November 13, 2025 at 7:10 PM
just tagging @tyrellturing.bsky.social
November 13, 2025 at 5:04 PM
just tagging @tyrellturing.bsky.social
One of the rare papers between superstar economist
@mioana.bsky.social and myself. This project is the result of trying to merge AI thinking with her econ approaches. There won't be a singularity. But neither will this be a nothingburger.
@mioana.bsky.social and myself. This project is the result of trying to merge AI thinking with her econ approaches. There won't be a singularity. But neither will this be a nothingburger.
November 13, 2025 at 4:42 PM
One of the rare papers between superstar economist
@mioana.bsky.social and myself. This project is the result of trying to merge AI thinking with her econ approaches. There won't be a singularity. But neither will this be a nothingburger.
@mioana.bsky.social and myself. This project is the result of trying to merge AI thinking with her econ approaches. There won't be a singularity. But neither will this be a nothingburger.
Lots of assumptions go into this. Play with it in the app we built: intelligencesaturation.org
2-Nest Economic Model with Endogenous Labor Allocation
intelligencesaturation.org
November 13, 2025 at 4:42 PM
Lots of assumptions go into this. Play with it in the app we built: intelligencesaturation.org
Early on in the transition, wages may go up as everyone becomes more efficient. As all intelligence tasks become automated, wages may drop. Overall productivity should go up. Profitability of AI companies should drop as saturation is approached.
November 13, 2025 at 4:42 PM
Early on in the transition, wages may go up as everyone becomes more efficient. As all intelligence tasks become automated, wages may drop. Overall productivity should go up. Profitability of AI companies should drop as saturation is approached.
The physical should quickly become the rate limiting component as AI becomes cheap and ubiquitous. In this sense our thoughts have deep links to existing ideas e.g. @dacemoglumit.bsky.social / Restrepo.
November 13, 2025 at 4:42 PM
The physical should quickly become the rate limiting component as AI becomes cheap and ubiquitous. In this sense our thoughts have deep links to existing ideas e.g. @dacemoglumit.bsky.social / Restrepo.
A task that requires physical inputs and intelligence inputs converges as the amount of intelligence goes to infinity while physical effectors are held constant. This changes our expectations for the future of the economy.
November 13, 2025 at 4:42 PM
A task that requires physical inputs and intelligence inputs converges as the amount of intelligence goes to infinity while physical effectors are held constant. This changes our expectations for the future of the economy.
In a new @brookings.edu paper we argue that we should not just divide into capital and labor, but also into intelligence and physical when modeling the economy (attention, economics speak): www.brookings.edu/articles/art...
(Artificial) Intelligence saturation and the future of work | Brookings
In a new working paper, UPenn's Konrad Kording and Ioana Marinescu present a novel framework for assessing AI and the future of work
www.brookings.edu
November 13, 2025 at 4:42 PM
In a new @brookings.edu paper we argue that we should not just divide into capital and labor, but also into intelligence and physical when modeling the economy (attention, economics speak): www.brookings.edu/articles/art...
Two very different systems! Optitrack is much more affordable
November 7, 2025 at 1:25 PM
Two very different systems! Optitrack is much more affordable
Client Challenge
pypi.org
November 6, 2025 at 9:06 PM
Matlab? Is it not supported at all by today's tools?
November 6, 2025 at 9:05 PM
Matlab? Is it not supported at all by today's tools?
Reposted by Konrad Kording
You will find this paper interesting and relevant too:
www.biorxiv.org/content/10.1...
www.biorxiv.org/content/10.1...
Misinterpreting the horseshoe effect in neuroscience
Dimensionality reduction methods are frequently used to analyze high-dimensional activity of cortical neuron populations during behavior. The resulting oscillatory trajectories that consistently emerg...
www.biorxiv.org
November 2, 2025 at 10:14 AM
You will find this paper interesting and relevant too:
www.biorxiv.org/content/10.1...
www.biorxiv.org/content/10.1...
There exist cases where circles are real. E.g. when cycling
October 31, 2025 at 5:58 PM
There exist cases where circles are real. E.g. when cycling
But also, jPCA exactly has the property of seeing circles when there are none.
October 31, 2025 at 3:26 PM
But also, jPCA exactly has the property of seeing circles when there are none.