I've been toying with LLMs for a while now, and here's my latest experiment: Synthetic podcasts with dead or fictional characters. Episode #1: Sherlock Holmes. We talk about Watson, drugs, physical exercise and role play.

Why is it so hard to explore these protein predictions!? Not anymore, since faltwerk v0.4 github.com/phiweger/faltwerk

Exploring an amino acid substitution in relation to ClinVar pathogenic variants and ligand binding domains in < 10 lines of code using github.com/phiweger/faltwerk

Thank you @martinhoelzer for pointing out the rabbit hole 😉 Also, here's an example of looking at multiple tracks of spatial information at the same time:

Have you seen the new Batman trailer? Yeah, but have YOU seen the new Bayesman trailer? youtube.com/watch?v=BYUykHSc…

What I did not know: Apparently you can now image the spatial organisation of the genome, too! #Plasman

listening to podcast by @lexfridman w/ @Plinz -- incredible pace of original ideas, currently re-listening. domains span synthetic intelligence, consciousness, cellular automata and dumb apes destroying the planet and, awesome youtube.com/watch?v=P-2P3MSZ…

@baym wouldn't the data from your (awesome!) macroplate experiment be a good ground truth for outbreak analysis methods? youtube.com/watch?v=plVk4NVI…