Профиль

Wait, doesn't capitalism make all broken things better?

Exactly. Leave them to it and leave the humans out of it.

maybe we can just leave all the exams and all the grading and all the university paperwork to AIs to fight out amongst themselves, and get back to just having discussions with students to teach them things if they want to learn.

In August, Jay Bhattacharya said “Training future biomedical scientists” was the 1st priority for his version of NIH. But talk is cheap. Let’s see how JB’s doing. 🤔 NIH supports trainees mostly via fellowship (F), training (T), and career development (K) awards. Here are funding curves for each.🧵

Just starting to read this paper on genome-wide convergence in fishes. Looks very worthwhile, and nice to see people using CSUBST. @agneeshbarua.bsky.social @msrivastava95.bsky.social Brice Beinsteiner Vincent Laudet @marcrr.bsky.social academic.oup.com/mbe/article/...

How did those axis rearrangements make it past the reviewers?

So the grad student was just foresightful in recognizing that you should have already written this? Sounds like you need to apologize for not having done so. \s

The question I wouldn't want to answer is how many papers were also there at the beginning of 2025, or beginning of 2024...

Yes, I agree with you that it is often not done.

Asking editors to do a semblance of their job seems kind of extremist these days, but is it too late to save peer review?

today I learned things about magic ponies that I never knew before

"Artificial Regurgitation" works, AR

Close cousin to "There are 10 programs out there to do this analysis, so we'll run all 10 and take the average. That's the ticket."

Midgely is generally a great read

I think if Vaughn reflected more on this being a true psychological and mental health imperative, he might be more accepting of having to ignore that inbox until January.

amen

That toxic waste will then pollute the data sets used to train the next generation of AI models, leading to ‘model collapse’ as the AI tries and fails to model incoherent garbage, vomiting out yet more incoherent garbage in the process. Meanwhile repositories of scholarly information are destroyed.

Right, why should 2026 get the credit for something that was mostly done in 2025.

PhD position available at the University of Bern in Prof. Claudia Bank's lab. Application deadline: Jan 9, 2026. Ideal for candidates with a strong math and programming background. More info: https://banklab.github.io/positions/ #phd

This looks nice. Thinking of going back to grad school so I can sign up...