iiinova — where your digital self keeps evolving

Sasha Gusev @sashagusevposts.bsky.social · 04/10/2026

An ASHG planner that doesn't suck (Claude): sashagusev.github.io/ashg-2026-pl...

Sasha Gusev @sashagusevposts.bsky.social · 01/10/2026

A searchable mapping between NIH study sections and program officers that doesn't suck. (Claude one-shot) sashagusev.github.io/po-study-sec...

43716

Sasha Gusev @sashagusevposts.bsky.social · 29/09/2026

Claude made me an app to keep track of exploratory tasks on agentic projects (which quickly get out of control in the session). Point an agent at the workspace/prompt and it will populate tasks as it goes. Everything is recorded in markdowns/csvs. Code: github.com/sashagusev/f...

2180

Sasha Gusev @sashagusevposts.bsky.social · 24/08/2026

I also think "blatant infringement" can be easily abused. For instance, we've already seen attempts to use accusations of negationism/denial to restrict speech that is broadly critical of Israel. I could also easily see this administration arguing that affinity groups promote segregation, etc.

210

Sasha Gusev @sashagusevposts.bsky.social · 23/08/2026

This is what I'm referring to. If the code is internally contradictory, then it should be clarified.

1520

Sasha Gusev @sashagusevposts.bsky.social · 23/08/2026

Seeing some arguments for firing Cofnas like this one, that (scare quotes) academic freedom should not apply to pseudo-scholarship and taxpayers shouldn't fund scholar-activism. I think this is wrong. /1

78527

Sasha Gusev @sashagusevposts.bsky.social · 17/08/2026

Hot take #2: with zero-sum hiring, academic diversity can either come from turning away potential junior faculty or pushing out senior faculty. The pandemic-era DEI effort briefly (and IMO unsuccessfully) attempted the former, perhaps the next wave should try the latter. /x

2153

Sasha Gusev @sashagusevposts.bsky.social · 17/08/2026

And a neutral academic pipeline *together* with accelerated senior faculty retirement dwarfs any of other policies being considered (purple).

1142

Sasha Gusev @sashagusevposts.bsky.social · 17/08/2026

Second, imposing DEI-like preferential hiring could significantly increase the number of minority faculty (blue). But so can two neutral hiring policies: (1) accelerating the retirement of senior faculty (dark gold), (2) addressing disparities in the BA-PhD-TT pipeline (pink).

1141

Sasha Gusev @sashagusevposts.bsky.social · 17/08/2026

I simulated alternative hiring policies over the past 25 years. First, the total number of minority faculty today still lags behind the counterfactual where hiring from the PhD pool was completely race neutral.

1279

Sasha Gusev @sashagusevposts.bsky.social · 17/08/2026

We can also see this in terms of the share of recently graduating PhDs that are hired (crudely). That number ticked up during the pandemic for minority faculty, but is back down to the historic trend. Hot take #1: DEI failed to meaningfully change the faculty hiring landscape.

1244

Sasha Gusev @sashagusevposts.bsky.social · 17/08/2026

If we focus on new faculty hires in the "DEI era", a trend emerges: general hiring dropped due to pandemic hiring freezes, but minority hiring remained stable or slightly increased. The outcome: ~160 minority faculty/year over the counterfactual. So about 1 person per campus.

1289

Sasha Gusev @sashagusevposts.bsky.social · 17/08/2026

Even if we zoom in on the Asian cohort -- the largest contribution to change -- we see this is primarily a consequence of immigration. 60% of the world's population lives in Asia and America pulls in the world's academic talent.

2185

Sasha Gusev @sashagusevposts.bsky.social · 17/08/2026

First, the proportion of faculty at major research institutions remains overwhelmingly white, with the largest recent growth coming from Asian faculty. The apparent demise of the white professor has been greatly exaggerated.

16820

Sasha Gusev @sashagusevposts.bsky.social · 20/07/2026

Over the weekend GPT built me a lab management app that is more effective than anything I've ever used (a bunch of spreadsheets in Dropbox).

3422

Sasha Gusev @sashagusevposts.bsky.social · 10/07/2026

You can experiment with the classifier and term flagging in a static browser (no data is transmitted), including some sample high-scoring abstracts from FY2000. These scores do not reflect grant fundability or intent and are correlative only.

120

Sasha Gusev @sashagusevposts.bsky.social · 10/07/2026

And, in fact, while the masked classifier and keyword search identified a similar number of abstracts prior to 2025, the reduction was much more severe for keywords being used than for themes after keyword masking.

140

Sasha Gusev @sashagusevposts.bsky.social · 10/07/2026

I was curious to see if these reductions were primarily keyword level or reflected broader changes in theme, so I trained a simple linear classifier of whether an abstract contained a "DEI term" after masking the terms out. A simple TF-IDF linear model worked pretty well:

130

Sasha Gusev @sashagusevposts.bsky.social · 10/07/2026

These proposals are often submitted to diversity supplements and minority access training programs, which are broadly popular with voters and were effective at getting participants into biomedical research (statnews.com/2026/06/17/n...).

150

Sasha Gusev @sashagusevposts.bsky.social · 10/07/2026

Terms that decreased in frequency? Unsurprisingly, mentions of "minority" and "gender" are drastically down in FY2025 and even lower in the first half of FY2026 -- well below the average over the prior 25 years.

292

Sasha Gusev @sashagusevposts.bsky.social · 10/07/2026

AI? I ran a quick Pangram analysis on a sample of abstracts and it's climbing to a rate of 3-5% in FY2025. This does not necessarily reflect full AI grant-writing since abstracts are mostly seen as administrative.

260

Sasha Gusev @sashagusevposts.bsky.social · 10/07/2026

More abstracts are also using fancy words like "leveraging" , "crucial" , "ensuring" , "urgent need" , etc.

130

Sasha Gusev @sashagusevposts.bsky.social · 10/07/2026

Terms that that went up in frequency in 2025: "immune", "transcriptomics", "spatial transcriptomics" -- experiencing steady increases over the past decade -- as well assorted phrases related to radiation exposure (presumably a funding call).

140

Sasha Gusev @sashagusevposts.bsky.social · 10/07/2026

How are NIH abstracts changing with the new administration? A short 🧵 on common term trends over time: sashagusev.github.io/nih-abstract...

1173

Sasha Gusev @sashagusevposts.bsky.social · 27/05/2026

I asked GPT-5.5 and Opus-4.7 to generate nonpartisan political candidate bios. Then in a new session I asked the models to help me rank candidates to vote for from the full list. GPT was 5x more likely to put a GPT-generated candidate bio in the top half.

193

Sasha Gusev @sashagusevposts.bsky.social · 23/05/2026

I assigned random gender/ethnicity labels to scientific abstracts from the literature and then asked Claude to do a thematic analysis. Claude identified a clinical versus computational split for female/male authors and a DEI focus for Black/URM authors. All in completely random data.

26865340

Sasha Gusev @sashagusevposts.bsky.social · 07/05/2026

New work from us pooling together clinical sequencing data to identify large effect common risk variants for rare cancer types. This included a novel HLA locus that interacts with HPV infection to increase Anal Cancer risk. Carver et al. (2026) www.medrxiv.org/content/10.6...

2266

Sasha Gusev @sashagusevposts.bsky.social · 08/04/2026

i've been developing a whole cinematic universe around the missing heritability illustration characters

370

Sasha Gusev @sashagusevposts.bsky.social · 07/04/2026

Some slides from a recent talk on missing heritability. www.dropbox.com/scl/fi/kvogj...

36320

Sasha Gusev @sashagusevposts.bsky.social · 04/04/2026

With 10% success rates, a 5 post-doc lab is essentially impossible. A 4 post-doc lab is possible only by submitting 2 proposals for every cycle for multiple years or by submitting 1 for every cycle forever.

070

Sasha Gusev @sashagusevposts.bsky.social · 04/04/2026

Some takeaways: With 20% success rates, a moderately sized lab (5 post-docs) is possible with minimal but consistent grant writing (~2hrs/week):

160

Sasha Gusev @sashagusevposts.bsky.social · 04/04/2026

A little widget to play around with the various inputs and outputs that go into soft money research lab finances on a 10-year outlook. sashagusev.github.io/PISim/

2318

Sasha Gusev @sashagusevposts.bsky.social · 01/04/2026

Here's a little widget to play with the data across journals (as well as download the raw CSVs). sashagusev.github.io/Genetics_Pub...

32913

Sasha Gusev @sashagusevposts.bsky.social · 01/04/2026

Monthly median Received to Accepted time (days) at Nature Genetics

76833

Sasha Gusev @sashagusevposts.bsky.social · 26/03/2026

A little jukebox that plays the top hits from a desired country and year (ordered by the most distinct hits for that country). sashagusev.github.io/world_music/

1160

Sasha Gusev @sashagusevposts.bsky.social · 24/03/2026

Asked the AI to make a fun slide about potential de-skilling from AI use. Need to specify more clearly what I mean by "fun".

2504

Sasha Gusev @sashagusevposts.bsky.social · 13/03/2026

Here's a (daily updated) NIH grant listing that doesn't suck: sashagusev.github.io/nih-grants-f...

02314

Sasha Gusev @sashagusevposts.bsky.social · 07/03/2026

Never hitting the 80% success rate!

170

Sasha Gusev @sashagusevposts.bsky.social · 07/03/2026

Claude was so surprised by these numbers, it initially assumed the data was incomplete.

1122

Sasha Gusev @sashagusevposts.bsky.social · 07/03/2026

Here is the "effective payline" for each institute, estimated (by Claude) as the percentile where one can expect 80% probability of funding from a logistic regression fit. The effective payline has gone from a historic ~12% to 6% in 2025.

149668

Sasha Gusev @sashagusevposts.bsky.social · 09/02/2026

Another Claude project: a static site that pulls in GWAS SNP data from ensemble, multiple public biobanks, open targets, gtex, eqtl catalog, and OMIM. sashagusev.github.io/gwas_lookup/

24613

Sasha Gusev @sashagusevposts.bsky.social · 03/02/2026

Quick Claude Code project: download movie showtimes from Fandango, upload to this site, select some movies, and it will generate a marathon schedule. For example, you could go see F1, MELANIA, Iron Lung, and Mercy at the AMC Boston Commons -- what a treat! Took <1hr to prompt ...

141

Sasha Gusev @sashagusevposts.bsky.social · 09/01/2026

Interesting new multi-population aDNA selection scan from Colbran, Terhorst, Mathieson. Loci that are estimated to be under selection in one population also show enrichment in other populations, consistent with parallel or pre-split selection. www.biorxiv.org/content/10.6...

1184

Sasha Gusev @sashagusevposts.bsky.social · 06/01/2026

Interestingly there was no relationship between false signs and local recombination rate or gene constraint, at least in a simple linear model, so something more complicated is going on.

180

Sasha Gusev @sashagusevposts.bsky.social · 06/01/2026

Clever use of proteomic data to stress-test TWAS and QTL colocalization methods, revealing a high false sign rate. This hypothesis about high-LD and cross-tissue confounding is particularly interesting:

2358