@milesalanmoore.bsky.social
PhD Student @ Univ. of Colorado Boulder (USA) | DOE CSGF ‘25 | mostly bayesian stuff, quantitative ecologist, computer scientist | plasticity, phenology, life history theory | connecting theory w/ data via complex models http://milesalanmoore.github.io
NO NO NO YOU DON’T
GUYS LET ME PLEASE SPREAD THE GOSPEL OF FREE TAX USA
federal is free, state is $16, handles even my chaotic freelancer taxes just fine, same step-by-step “designed for normies” kind of interface as TurboTax but NOT EVIL
tell everyone you know
www.freetaxusa.com
How your email finds me
New postdoc opportunity in our lab!
good morning everyone project your personal imposter syndrome onto this gif ur welcome
I'm too depressed to dunk on the constant stream of bad statistics takes on the site. If, however, you're looking to lurk through or engage in productive discussions on (narratively generative) probabilistic modeling then drop by the Discord server I set up, www.patreon.com/posts/genera....
Adios by Antonio Saura, 1959
https://botfrens.com/collections/212/contents/137672
world-altering decisions are being made by a cabal of gambling addicts hell bent on self enrichment no matter the cost to humanity, we should look into that.
Im not sure how I feel about Marathon but I definitely agree with my man Peanut about how players continue to discourage new IPs.
youtube.com/shorts/2S6yH...
One of the most ridiculous saves I have seen for a while. Doing that at Olympic final makes it even extra ridicuolus. We have a thriller !
Defend democracy like Connor Hellebuyck defended goal
Unreal stick save by Hellebuyck on Toews
This really is a great picture:
A personified AI agent submitted a github PR to matplotlib. The project maintainers closed the issue, then the AI (really, the developer controlling it) published a blog post on "its" blog accusing the project maintainers of discrimination, then the LLM (again, the developer) had to issue an apology
In the Dune universe, only the Butlerian Jihad was able to eliminate infinite scrolling
This isn't shortread data, but alignments of genome assemblies with N50 > 100Mb. The plot is the # of bp per 100Kb window that are dropped because of missing data -- because those bp physically are not there in one of the other 25 genomes. This is what normal plant genomes look like, folks.
Susan Jocelyn Bell Burnell (1943-) is a Northern Irish physicist who, while conducting research for her doctorate, discovered the first radio pulsars in 1967. This discovery earned the Nobel Prize in Physics in 1974, but not to her - the prize was awarded to her (male) supervisor. #CelebratingWomen
Anything you can do in Obsidian you can do from the command line.
Obsidian CLI is now available in 1.12 (early access).
Water Lilies isn’t one painting—it’s a series.
Monet returned to the subject again and again, creating over 250 canvases. One pond, endless light.
as I'm revising my course materials, I keep stumbling upon cool @mc-stan.org developments.
Current favorites:
1. your model has funnels and you exhausted reparametrization ideas: metric = "dense_e" makes your HMC learn about covariance btw parameters. Sloooow, but effective!
1/
New pre-pre-print:
Finally listened to this lecture -- highly recommended if you've ever been confused about fixed effects vs. random effects models. Love that the framing starts from group-level confounding -- here's a problem, what could be solutions (rather than: here's a model, what are its features).
Video shows ICE leaving behind an infant and broken glass after arresting a man with no criminal history.
“There was a car seat in the back... There were broken glass shards all over it... There was just this tiny peanut of a baby. He was crying.”
My latest: www.pressherald.com/2026/01/28/i...
I guess he just got lucky.
Data centers are coming to Colorado. Can the parched state handle their large-scale water needs?
“Fear not the those without papers, but those without conscious.”
www.instagram.com/p/DT8OwTxjSbm/
Lecture B02 of Statistical Rethinking 2026. Multilevel model expansion: workflow, cluster and feature engineering, non-centered parameterization, dank memes. Continues next week with models of population covariance and (if there is time) group-level confounding and "Mundlak machines".
The trick here is basically to run screaming away from dichotomizing findings and into the warm and loving embrace of the vast language we have for describing uncertainty and scale, interpreting whether it matters accordingly. Use figures that make it clear.
In my mind, it’s very often the case that one can cast a frequentist model as a special case of an analogous Bayesian model (albeit not without simplifying and handwaving in both directions). I guess this can be somewhat useful pedagogically but not a helpful heuristic in general?
Simulate, then fit. Simulate, then fit. Simulate, then fit.
Please share: looking for postdocs to work on the theory/ concepts of multiple factor impacts.