RT Raghavan
@rtraghavan.bsky.social
Engineer formerly @NYU, Duke. Working in computational vision. Tweets are my own
My first thought, someone messed up the Cuda update.
December 27, 2024 at 10:16 PM
My first thought, someone messed up the Cuda update.
Marr discourse among neuroscientists is like people who quote Adam smith but have read *at most* one chapter of wealth of nations.
A few weeks ago Sebastian Seung (are you here yet?) posted
"Marr's three levels are dead. AI killed them."
and a fun discussion ensued.
But although many of us grew up with the 3 levels, it was not clear to me that anyone actually still believes in and defends them?
Is this still debated?
"Marr's three levels are dead. AI killed them."
and a fun discussion ensued.
But although many of us grew up with the 3 levels, it was not clear to me that anyone actually still believes in and defends them?
Is this still debated?
December 1, 2024 at 4:05 PM
Marr discourse among neuroscientists is like people who quote Adam smith but have read *at most* one chapter of wealth of nations.
The shame is that neither option describes the majority of the actual work people do, which is usually filling in details, refining measurements, etc. So we end up with a framing that makes it harder to distinguish signal from noise
There are basically 2 ways to make a splash in science.
1). The current paradigm is wrong
2). Here’s something you didn’t think of.
Both are valuable. Different people veer to each side, but the first set of people end up in more arguments:)
1). The current paradigm is wrong
2). Here’s something you didn’t think of.
Both are valuable. Different people veer to each side, but the first set of people end up in more arguments:)
confession: i was rarely driven by curiosity. i was driven by arguments. usually i saw something wrong & wanted to make the point. i did experiments to bring more ammunition for the next round of debates.
isn't that a bit sick? should i stop? would i still be ok if i just try to make things work?
isn't that a bit sick? should i stop? would i still be ok if i just try to make things work?
November 27, 2024 at 7:40 PM
The shame is that neither option describes the majority of the actual work people do, which is usually filling in details, refining measurements, etc. So we end up with a framing that makes it harder to distinguish signal from noise
Reposted by RT Raghavan
I am old so I can say this: innovation in neuroscience research can be accelerated if we did not waste so much resources repeating what we already know from NHP and rat research in Cre mouse lines
November 25, 2024 at 6:14 PM
I am old so I can say this: innovation in neuroscience research can be accelerated if we did not waste so much resources repeating what we already know from NHP and rat research in Cre mouse lines