Dan F-M
banner
dfm.io
Dan F-M
@dfm.io
Retired computational astrophysicist
Interpretability at Anthropic; prev Google DeepMind, Simons Foundation, Journal of Open Source Software
(he/him)

https://dfm.io // https://github.com/dfm
Thank you!!
July 15, 2025 at 10:25 PM
Don’t worry - we’re still repping open publishing models over here… 😍
July 15, 2025 at 6:59 PM
Given the groans that echo through the entire astronomical community every time it hits the news again, I'm not sure this is a fact that needs to be publicized… 🫠
April 20, 2025 at 1:15 AM
So bad! We haven't seen the "baby loves" books yet, so I'm glad to hear that there are some better options in the genre 😀
December 10, 2024 at 12:19 PM
The values there are actually used as an initial guess for the optimization. It is circular, but in any real world application you'll also need to provide decent (whatever that means :D) initial guesses.

I mostly point people to tinygp.readthedocs.io instead of george but either should be ok.
tinygp — tinygp
tinygp.readthedocs.io
July 20, 2024 at 4:56 PM
Actually, I lied. You can use:

from numpy import sin as sine

but (I think?) not

import numpy.sin as sine

So there is a difference in that direction. But for modules they should be the same!
July 16, 2024 at 8:45 PM
I don't think there is any difference! I certainly use the two interchangeably and I'm having trouble thinking up potential pitfalls…
July 16, 2024 at 8:42 PM
I once (many years ago) followed a recipe for pesto which included "3 garlics", and I said "3 full garlics is way too much… let's start with just 1!" It was almost edible.
December 18, 2023 at 9:30 PM