David Pollock

@genomesevolve.bsky.social

Evolutionary and Mathematical Biology, sequence structure and function, genomes evolve, statistical theory and knowledge, empirical-theoretical interface, communication of evolutionary genetics, collaboratives, the mind-body problem, humanism

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.🧵

Bild

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.

This is an interesting distraction driven by @jessicacalarco.com. The comment below misses the point that human-support (and art) wage rises are a lagging effect. It seems that what needs to happen is to accelerate the effect to counteract inequality. UBI perhaps? Decrease educational costs?

Paul Hünermund@p-hunermund.com · 8mo ago

The Baumol effect leads to rising wages in jobs with low productivity growth (such as dog walking) and thus counteracts inequality across sectors, by the way.

recent protein language models can extract embeddings from protein sequences reflecting high-order features, developed statistical tests to evaluate the adaptive convergence of such features. Interested to see how people will use this (I edited it, not an author). www.pnas.org/doi/10.1073/...

Language models reveal a complex sequence basis for adaptive convergent evolution of protein functions | PNAS

Convergent evolution, or convergence, refers to repeated, independent emergences of the same trait in two or more lineages of species during evolut...

pnas.org