Genes, Minds, Machines

Protein language models are overly constrained by covariation

Why good performance on one set of goals can lead to poor performance on another

Aug 12 • Claus Wilke

16

6

1

The error bar cargo cult

You don't need to understand error bars. You just need to draw them.

Aug 2 • Claus Wilke

46

10

7

July 2026

Nene Royal—A generational talent

The kids in Thailand are alright.

Jul 30 • Claus Wilke

11

3

1

If the work is good why does it need a story?

I thought we do science here, not fiction or romance

Jul 21 • Claus Wilke

26

9

9

How useful are zero-shot predictions of mutational effects?

Zero-shot predictions are great, except for when it matters

Jul 13 • Claus Wilke

17

3

1

“Tell a story” is not helpful advice

If people knew how to tell a story they wouldn't need the advice in the first place

Jul 6 • Claus Wilke

29

4

2

March 2026

The Google Scholar preprint bug strikes again

Google is never going to fix this bug, are they?

Mar 31 • Claus Wilke

8

Creating reproducible data analysis pipelines

There was a discussion recently on Bluesky about reproducible data analysis pipelines.

Mar 27 • Claus Wilke

44

5

4

Protein language models are bad at mutational effect prediction

Biology is hard. Yes, even for AI.

Mar 19 • Claus Wilke

22

7

2

February 2026

Sociopathic AI agents

AI alignment will likely require creating AIs with genuine empathy

Feb 15 • Claus Wilke

15

3

3

November 2025

Python is not a great language for data science. Part 2: Language features

It may be a good language for data science, but it’s not a great one.

Nov 17, 2025 • Claus Wilke

37

17

6

Python is not a great language for data science. Part 1: The experience

It may be a good language for data science, but it’s not a great one.

Nov 13, 2025 • Claus Wilke

91

44

11