Michael Waskom

come work on some of the hardest problems in AI @modal

Rand Paul: Do you realize that the man that beat Charles Sumner with a cane, he beat him till he was unconscious. That's what you're insinuating, is the precedent of the senate, and that's what you live by. That is a very, very dangerous sentiment.

I'm hiring for a new role @modal: Agent DX Researcher Most of the code deployed on Modal will be written by agents. We're getting scientific about optimizing agent productivity and providing the best possible DX for building AI applications. Join us! jobs.ashbyhq.com/modal/407cd498…

@iam_lpettingill We aim to make the modal American experience fast cold starts and seamless autoscaling

@shoyer @exoplaneteer @FlatironCCA @GoogleDeepMind Congrats in both directions!

@exoplaneteer @FlatironCCA @GoogleDeepMind So excited to have you onboard at Google!

Some news: in a few weeks I'll be leaving @FlatironCCA to join @GoogleDeepMind where I'll be working on the JAX core dev team. It's bittersweet to leave astro which (for better or worse) has defined much of my identity for the past 15 years, but I'm excited about new adventures!

@patrickmineault Will it be much work?

Be a content creator for the Neuromatch NeuroAI course! We're looking for people to write tutorials on transfer learning and RL in PyTorch over the next 3 weeks. If you want to help build this amazing course, DM me for details or fill out this app: airtable.com/app32npl2ZlbJv…

Some professional news: excited to say that I’ve joined @modal_labs to help build a completely new approach for developing data and AI applications in the cloud!

My talk on the past and future of seaborn from #positconf2023 is now available on YouTube youtu.be/JE2C1MhZO6E

Logistic regression is AI when it’s regularized by default ;)

This is a well targeted ad in the sense that I also mostly use my Mac to make @matplotlib plots

@fperez_org It’s interesting to look back at the example notebook linked there, which has many of the core ideas almost as they are today nbviewer.org/github/mwaskom…

But then it got a little blip of attention on twitter from influential people like @fperez_org and a release felt harder to avoid (seaborn is so old, from the age of manual quote tweets) twitter.com/michaelwaskom/…

Apparently it is ten years today since the first release of seaborn on PyPI 🎂

Maybe I'm old fashioned, but I prefer my prices quoted in dollars instead of "Regular Twitch Neurons" and "Fast Twitch Neurons". Idk maybe just me?

@shoyer Then I’m not understanding how you’re ending up with different size fonts than you set?

@michaelwaskom Usually PDF

Any favorite tricks for making figures with the right font/line sizes for scientific papers in matplotlib? My attempts to modify default font sizes (e.g., seaborn.set_context) have not worked well.

@shoyer What file type are you saving to?

I am currently favoring "scale down by 2x" where I set figsize twice as big it should be and then shrink by 50% when inserting figures into my manuscript. In theory this should come close for adjusting font sizes from matplotlib defaults (12pt) to publications sizes (~6pt)

@choldgraf I’m never unpinning pydata-sphinx-theme=0.10.0rc2

@michaelwaskom import seaborn as plt ✅

There’s a fresh new release of seaborn available on PyPI: pip install seaborn==0.13.0 New features include some major improvements to categorical plots, support for alternate dataframe libraries, default theme/display configuration for so.Plot, and lots more.

@choldgraf Chris.

@michaelwaskom Wait the preferred abbreviation is now so and not sn?

For the full set of release notes, see the docs: seaborn.pydata.org/whatsnew/v0.13…