davidaknowles.bsky.social
machine learning and functional/statistical genetics. Assistant Prof @Columbia and Core Faculty @nygenome. he/him/his. https://daklab.github.io/
116 posts
2,881 followers
507 following
Prolific Poster
Conversation Starter
comment in response to
post
Just anti funding any science, completely different!
comment in response to
post
Thinking back on all this I better understand the pain I feel to see science under devastating attack here. It’s not just about my livelihood or my university. It’s about my identity. And it’s about a pursuit that I see as standing along with art, literature, and music as among our highest callings.
comment in response to
post
It's been very helpful around NYGC in Soho, no more stationary honking cars trying to squeeze into the Holland tunnel.
comment in response to
post
Just explain Bayesian credible intervals instead!
comment in response to
post
python/torch, all fp32. it's a floating point (and order of operation) issue, so could happen in any language.
comment in response to
post
Curious if any of the LLMs would have spotted this as a potential gotcha.
comment in response to
post
Refried beans go straight to my knees.
comment in response to
post
Sounds like an excellent pairing!
comment in response to
post
You want to know recall, but we need to know what precision you're OK with.
comment in response to
post
UPDATE: Nature.com has issued a correction: “Correction 14 November 2024: An earlier version of this article stated that AlphaFold3 is now ‘open source’. This is not the case, as weights for the model can only be accessed upon request, and the code can only be used for non-commercial applications.”
comment in response to
post
I always prefer weighted graphs.
comment in response to
post
Important to remember this for authorship. The previous lab member who generated data/did analysis that didn't make it into the final paper, but helped rule out dead ends, should be an author.
comment in response to
post
Somatic mutation I'm thinking.