Neuroscientist. We aim to understd brain networks and brain wiring, using lasers & neuro AI. Prev: policy for @democracypolicy. Personal views. Not a lab acct.

Joined May 2013
411 Photos and videos
Pinned Tweet
11 Mar 2024
New preprint, on 'sequence filtering'. Led by @CianaDeveau, Z Zhou. We see this as a key step forw on how cortex works. All cortical areas have dense exc-exc recurrent connectivity. What do these connections do, esp in sensory ctx? Our data say: they do dynamics/time. 1/4
6
39
176
35,099
"position does not confer appeal rights to the MSPB... prohibited personnel practices will [not be enforced by] OSC" We're seeing them go straight for the Civil Service Reform Act and CSRA appeals. And the Roberts Court has been signaling they're good. Cool times! (sarcastic)
1
338
chat, is destroying an independent, expertise-driven civil service a good thing for US science, innovation, and harnessing the US talent pool?
1
182
Amongst a long long literature on this: " Exploiting the implementation of the Pendleton Act–a landmark statute which shielded bureaucrats from political interference, we find that civil service reform reduced delivery errors and increased productivity." guoxu.org/docs/SSC_Postal_An…

36
A foundation of NIH is peer review. Rotating panels of external scientists hugely influence which projects get funded. As it shld be, b/c scientists, not politicians, are best-placed to choose the most promising science. Changes to peer review… 1/ 27unihted.substack.com/p/the…

1
3
529
threaten to collapse NIH’s success. Also, the rotating nature of peer review panels engages with a large section of the nationwide scientific community. This distributed decision making resists political capture. Keep a close eye on ongoing changes to peer review.
1
391
Mark Histed retweeted
Yep. Check out our 2014 paper on this: papers.nips.cc/paper_files/p…

Local minima are rare in high dimensions because a strict local minimum has to curve upward in every direction, so all Hessian eigenvalues must be positive. In a D-dimensional toy model where eigenvalue signs are independent, that’s a 2^(-D) event. In GOE-like random matrix models, positive definiteness is even rarer, roughly exp(-cD^2). So as dimension grows, random critical points are much more likely to be saddles than minima. This is one reason high-dimensional optimization is often a saddle-escape problem, not a bad-local-minimum problem. Wrote up some of the math here: grantstenger.com/local-minim…
4
17
195
32,032
TPM covers the presidentialization of science agencies. This is what Natalie Aviles and I have been on about: the unitary executive logic of SCOTUS/P2025 is not compatible with good science. "[the OMB rule] purports to end the system which Congress created..." 1/
1
375
"...in which the fed gov't funds sciences and medical research. Instead, it makes the entire apparatus into a kind of political patronage system, with any researcher liable to be immediately cut off if they offend the current occupant of the White House." talkingpointsmemo.com/edblog…
1
355
"This isn’t how the sciences work, and it’s basically a dagger aimed at the sciences and medical research" That's right
240
“As the admin moves from cutting budgets and attacking elite universities to dismantling the foundations of the US research enterprise, pharma CEOs have remained silent. Their silence is a choice.” biocentury.com/article/65961…
2
2
9
1,835
“The damage inflicted over the past 16 months is prologue. The OMB’s proposal for overhauling U.S. government grantmaking would explicitly subordinate expert peer review to political gatekeepers and make ideological alignment with the administration a prerequisite for funding.”
1
70
Hopefully this moves the needle and encourages the pharma industry to speak up. I have heard from several sources that pharma execs know exactly what’s going on and what the damage will be, but they can justify inaction because the effects won’t hit their stock prices soon.
46
I deeply believe in a non-partisan civil service. But civil servants must also be pro-democracy and anti-fascism. If that runs at odds with the choices of the political parties today, that is not the fault of civil servants.
1
1
1
518
And civil servants should not bend their values. They should not give up their commitments to democracy.
1
2
457
Governments in personalist states don’t work. And moreover civil servants in the US took an oath to the Constitution— that is, to American democracy. If one party chooses to shred the Constitution, that is also not the fault of civil servants.
183
High-risk, high rigor, and low cost in science are tradeoffs: for any line of research we generally get to choose just two. For some emerging fields, we want to throw a lot of high-risk and low-cost darts …
1
1
4
678
and if any one is wrong we can accept it. In contrast, some big clinical trials maximize rigor at high cost. A great science ecosystem has a mix of different approaches. #metascience
1
123