gammarator

he just...he tweeted it out

At the present moment I think science is way more threatened by the OMB absconding with the grant budget than it is by AI.

“Just outside of Boston,” you say? They were coming to get Avi Loeb.

Again, the status quo is that funding recommendations are made by expert peer review by PhD scientists. Political appointees literally do not have the knowledge to make the calls this policy directs them to.

the bureaucrats that were making the decisions before were not

Russ Vought was not elected either.

Here's a more concise summary of the proposed changes: https://elizabethginexi.substack.com/p/summary-of-key-change...

I don't think any practicing scientist of any political persuasion will think these are good for science.

Science progresses by sharing knowledge openly and publicly, so others can evaluate it, criticize it, and build on it. These severe restrictions on collaboration, publication, and public communication will damage science's naturally open, merit-based culture.

We will all suffer due to lost discoveries--maybe not today, but over years and decades.

"R&D" is not the same as "grants supporting fundamental science."

You are misreading how Broader Impacts (BI) works. From your link:

Some examples that illustrate contributions in each of the five areas are given below. Proposals need not address all of these areas, and PIs are advised to focus on those areas in which they are well prepared to make meaningful contributions.

"Broadening participation of underrepresented groups" is only one of the five areas, and no proposal was required to use it. I had proposals funded that focused on workforce development, for example. I saw others focus on science communication to the public (now forbidden in the memo this post is about!).

Proposals that passed grant panels were first and foremost always those that would great science. At ~10:1 oversubscription rates or more, proposals don't pass without it. The BI component needed to be credible but could be handled lots of ways.

Fundamentally, Congress recognized when defining BI as a component for merit review in the NSF that fundamental science only pays off in the long term. BI is a pragmatic choice to ensure that grants also yield near-term benefits to society as well.

The political system, representing the taxpayer (primarily via Congress), has always dictated scientific strategy--do we build the Superconducting Supercollider or cancel it; do we return a sample from Mars or not; do we sequence the human genome. How big a budget do we devote to medical research compared to physics, etc. Scientists advocate, but politicians decide.

However, the nuts-and-bolts day-to-day tactical decisions have before been made through expert peer review, by scientists. Given a fixed and finite budget set by Congress, what is the best way to make discoveries?

Having been on grant review panels, it's brutal--at 5 or 10:1 oversubscription rates, your peers will find any flaw in your proposal.

Political appointees are deeply unqualified to make these judgments. To take a very specific example: the White House Office of Science and Technology Policy is headed by Michael Kratsios. He has a BA in politics and has never written a scientific publication. (Every prior OSTP head was a PhD scientist.) The OMB memo says he and those like him should decide what to fund without deferring to scientists. How is he going to assess which of 50 proposals on (say) "hypothalamic SH2B1 neurocircuits and SH2B1 signal transduction pathways" is the good one?

He can't, so he'll either choose AI or graft. Both are destructive to our once world-leading scientific enterprise.

the text of this piece is entirely human-generated, including the infelicitous phrasings and penchant for two-dollar words.

“I aim for the stars… but sometimes I hit London.”

“and, like, sort of ethics”

I have a 20 year old Malm dresser that made four moves and is doing just fine.

(I do have it screwed into the wall as it’s been recalled for tipovers.)

You’ve got to normalize by dollars spent, though.

Ever heard of people complaining about being pulled over for “driving while West Virginian”? Why or why not?

Oh look! https://garymarcus.substack.com/p/the-whole-thing-was-scam

I would not be surprised if Sam A. helped engineer this whole situation… “Child’s play,” like replacing a reddit ceo.

Did you read the article? The author considers these possibilities and offers their estimates of the odds of each. It’s fine if yours differ but you should justify them.

Also “The Unaccountability Machine” https://press.uchicago.edu/ucp/books/book/chicago/U/bo252799...

Evaporative cooling of the best content, likely, but plenty of social networks with worse ad loads still have plenty of users.

My question is whether Ellison does to it what Elon did to X: revamp the algorithm to support his politics.

This is not very newsworthy, stellar flares are common and known to emit X-rays.

Sounds like Dr. Kirkpatrick should have been a coauthor.

While Pasadena is a relatively wealthy city, historically there has been significant avoidance of its public schools by affluent residents: https://southerneducation.org/in-the-news/new-polling-data-f...

The sentence is accurate, your comment is not. The administration unilaterally canceled existing grants and halted and showed granting of funds appropriated by Congress, so the money was not used as allocated. If Congress allocated $1B for medical research and the administration only releases funds for $500M, it’s ignoring the law.

For example, https://www.nature.com/articles/d41586-025-01396-2

As to “exercising control,” American science has been great because scientists judge which projects are the strongest. That’s being replaced by judgement by political appointees who are not experts: https://www.science.org/content/article/nsf-pares-down-grant...

Your linked article does say Rivian ranked first in satisfaction, which does support the GP’s “rabidly loyal.”

Can’t wait for the brave new world of individuals “match fixing” outcomes on Polymarket.

(American _Astronautical_ Society)

[citation needed]

Emacs and vim have complex interfaces that have been stable for decades. Seems like every new flavor of LLM requires learning its warts and blind spots from scratch.

Lobby groups are private actors rating elected Congressional reps. This is government officials rating private actors (with an implied threat to punish those who don’t comply).