William Thompson

Announcing Octofitter (🐙), an ambitious tool for all kinds of exoplanet orbit modeling workflows: fit planet orbits to relative astrometry, proper motion anomaly, radial velocity, images, and more! [1/7] sefffal.github.io/Octofitter.jl/…

fun fact: the human body contains 400 AU of DNA

Great work by Sabina and co-authors! Plus lots of nice showcases of Octofitter and PairPlots :)

New paper! In work led by Sabina Sagynbayeva we quantify mission and ancillary data (RV, astrometry) requirements that enable "full system" (terrestrial AND giant planets) characterization with the Habitable Worlds Observatory. arxiv.org/abs/2507.21443

@wsmoses @MilesCranmer Great to hear! I’ll make sure to test out the latest versions. I will say that those issues aside, the performance from Enzyme seems totally unmatched by any other package, so I do use it personally even when I haven’t yet recommended it to users directly.

@AstroWrt @MilesCranmer Unfortunately a lot of those errors are out of Enzyme.jl's control (and often error's in Julia's GC itself that are only hit for fast code, indep of Enzyme). That said we got all that I know of fixed in the latest releases so I'm curious if you still hit any

I am a HUGE fan of Mooncake.jl github.com/chalk-lab/Moon…, a new language-level autodiff for Julia. Most impressive is how uncompromising the devs are about correctness and reliability. It has some of the best software engineering practices I’ve seen in the entire ecosystem!

@MilesCranmer This has been my experience with Enzyme too! I can get it working well for myself but I’m not sure I trust it enough to roll out to others.

@AstroWrt It's much wider coverage than Zygote though, which can't differentiate in-place ops. Enzyme is probably the most general. Though I always seem to generate new segfaults in Enzyme (likely from my own misuse!) so I've never been able to use it reliably.

A surprising & little-known results in classical statistics: Mean (μ) and median (m) are within one std deviation: |μ−m| ≤ σ For unimodal densities, bound is even tighter |μ−m| ≤ 0.7746 σ This beautiful results first appeared in a 1932 paper by Hotelling & Solomons 1/3

@ksm_wilson From the Fairmont too, from the looks of it 👀.

Not a bad view for an evening in the metrop.

Our paper on a differentiable spectral model, ExoJAX2, has been accepted by ApJ. In this paper, we use HMC-NUTS to analyze JWST’s native-resolution transmission, high-dispersion spectra of a brown dwarf from Subaru, and Jupiter’s reflected light. arxiv.org/abs/2410.06900 1/N

@MonaMounsa @AstroCanada @PHASTatUVIC Thank you!

@AstroWrt @AstroCanada @PHASTatUVIC Congrats 👏

Just learned that am being awarded the J. S. Plaskett medal for most outstanding doctoral thesis in astronomy in Canada. Thank you @AstroCanada, and my committee & @PHASTatUVIC for the nomination!

@djulik @AstroCanada @PHASTatUVIC Thanks!!

@AstroWrt @AstroCanada @PHASTatUVIC Well deserved, congrats!!

@ksm_wilson @AstroCanada @PHASTatUVIC Thank you! Looking forward to it!

@AstroWrt @AstroCanada @PHASTatUVIC Heartiest congratulations, sir! Looking forward to celebrating this weekend!

@AstroCanada @PHASTatUVIC

Taking my glasses off when I get home after a long day because I’ve seen enough

The Flag That Unites Us. Happy #FlagDay! 🇨🇦

Researchers from UVic’s Astronomy Research Centre used the James Webb Space Telescope to captured a rare glimpse of how young planets are forming. This groundbreaking discovery reveals how planets compete with their host star for material shorturl.at/Ei7Fh @ArcUvic #JWST

Just saw someone saying using the em-dash is a sign of something being written by AI, because real people rarely use it. This is terrible news for everyone like me who has an unhealthy emotional attachment to the em-dash.

Load in 4k to spot the ISS