Thursday Links

1. “Brain-computer interface company Precision Neuroscience says that it has set a new world record for the number of neuron-tapping electrodes placed on a living human's brain—4,096…”

2. Several important CRISPR papers.

   1. This paper mapped (accidental) chromosome truncations caused by gene-editing.

   2. This paper solved protein structures for a prime editor to figure out how they work in greater mechanistic detail.

   3. Type II CRISPR systems have trans-activity, meaning they can cut single-stranded DNA or RNA separate from the target strand.

3. A microbe called Cloacibacterium sp. CB-01 “reduced N2O emissions by 50–95%, depending on soil type” in field trials. And it’s not engineered, which is good news for adoption.

4. $100-$500 microgrants for “ambitious young builders.” See also Fast Compute Grants.

5. Mistral AI releases Codestral, an open-weight code model trained on 80+ programming languages. Also: The amount of compute used to train frontier models is scaling by ~4.2x/year.

6. They’re made out of meat.

7. Interview with Michael Nielsen. Recommend.

COWEN: Now open science — why do some fields have preprint platforms and others not? Is there an actual regularity, or is that random and path-dependent?

NIELSEN: I think a lot of that probably comes down to individuals. One of my favorite things, years ago, before they’d started to spread in biology, I would often ask physicists and biologists this question: “Why are there preprints in physics but not in biology?” The biologists would say, “Biology is so much more competitive than physics that we can’t possibly bear to share our results too early.” The physicists would say, “Physics is so much more competitive than biology that we have to share them as rapidly as possible to get the word out.”