Josh Peek 🌴 πŸ₯₯ (@jegpeek) 's Twitter Profile
Josh Peek 🌴 πŸ₯₯

@jegpeek

Associate Astronomer and Head of Data Science at @stsci. Principal Investigator @MAST_News. Research Scientist at @JohnsHopkins. Occasional tweets all my own.

ID: 20697617

linkhttp://jegpeek.space calendar_today12-02-2009 17:52:24

15,15K Tweet

1,1K Followers

606 Following

John F. Wu (@jwuphysics@bsky.social) (@jwuphysics) 's Twitter Profile Photo

Not to be all "us vs them," but this is exactly what our team has been trying to avoid. - no human evaluations - generates nonsense figs/tables - LLM "peer review" Ensuring that LLMs are actually useful for science isn't the main thing, it's the only thing.

Josh Peek 🌴 πŸ₯₯ (@jegpeek) 's Twitter Profile Photo

I have done this, twice. Once to understand the analemma, once to figure out a sky coordinate transform for an app I was writing. Would recommend.

Josh Peek 🌴 πŸ₯₯ (@jegpeek) 's Twitter Profile Photo

Hi folks! I've now come across just straightforward antisemitismn (not anything to do with zionism / israel / gaza) scrolling X twice in two days now. See you on bsky. 🫑

Grant Tremblay (@astrogrant) 's Twitter Profile Photo

Of course there are countless ways to cut costs and increase efficiency, but there is vastly more waste than fraud. Example: in order to spend $500 of taxpayer money, my org must spend $5,000 in integrated Admin FTE time in order to prove that it's *not* fraud.