This was done with @anthropic.com's Claude Code. You can grab the code yourself here: github.com/rtfisher/tor... .

Definitely one way to run out of Gauss.

Definitely a huge challenge but I know we will eventually get there. It's something I also confronted when teaching math methods; software systems have been better at doing integrals than introductory students for decades, but that doesn't mean we have stopped teaching calculus.

They are extremely useful for many tasks -- from generating comments to code or git commits to developing perfectly formatted figures. But they are improving so rapidly we cannot really ignore them. A new generation of agentic tools like Claude Code are already here.

¡Todavía tengo el vinilo original!

Hope the historic Orrery made the cut!

This looks cool!! The postdocs and students at @hitsters.bsky.social developed a Habitable board game last year while I was there. There was a limited run of physical copies, but there is a version available online for free. tabletopia.com/games/habita...

When I was a teaching assistant for Alex Filippenko’s large introductory astronomy course, Alex solicited us (the TAs) for exam questions. The end product was some combination of the best questions from both Alex and the TAs.

Pre-tenure: ~ 1 course per semester. Post-tenure: ~ 1 course / 5 years.

I usually do this as a homework assignment, but it could work in-class: given stellar data, “discover” the HR diagram by plotting mags and then abs mags versus colors. Also interactively see the difference between a magnitude and volume limited sample.

Indeed, we had a preview of what a sudden endowment shortfall looks like — Harvard for instance couldn’t even afford to put breakfast on the table for their students. That is precisely their goal here.

Very cool, thanks!

Is this a Seestar?

There's actually quite a literature surrounding hypothetical "neutrino balls," going back at least to Holdom, 1987 (journals.aps.org/prd/abstract...).