@deanwball The "ungovernability" (and openness) of Linux and the Internet is precisely what has made their success.
The same will be true of open weight AI foundation models.
Friends,
Sara is recovering well from surgery. Her cancer has progressed to a higher stage and needs more aggressive treatments. Please keep sharing/donating to help us make these crucial treatments possible, and thank you deeply for your love and support! https://t.co/AQlBFvAy1g
Excited to announce my first-ever paper!
We show that the predictive coding principle (as implemented in PredNet), previously suggested as a candidate mechanism for peripheral drift illusions, does not reliably give rise to human-like illusory motion.
https://t.co/WJy41qAW3q
Super proud to have my first *last-author* paper, with @Kirubes1 who has done an amazing job pushing this to completion over the 1.5 years we ended up working together.
Just checked, and it's 11 years after my own first-ever paper came out, also in Vision Research!
Future work on peripheral drift illusions in DNNs should take our findings as a note of caution and revisit the question of whether there are architectures and training regimes that do robustly produce such illusions. 7/