Lmst

Nice: my CV just hit 69 publications

@hattom Haha you know my mindset right! Just skipped the ones where people broke the paper build

Bonus: here's an animation I generated showing how the sausage was made. Each frame is one commit from the paper repo.

Something my SXS colleagues and I have been working on for the past six years: https://arxiv.org/abs/2505.13378

The SXS Collaboration's third catalog of binary black hole simulations
Scheel et al.

We have a little news blurb here: https://www.black-holes.org/2025/05/19/catalog-update

Here's a nice figure — read the news blurb for a short explainer, or the paper for more details!

Figure 1 from our new catalog paper. We accurately capture precession, memory, eccentricity, and high mass ratio systems. For full details, see the paper.

🎉 arXiv submission day! 🎉

🎉 Paper acceptance day! 🎉

I was Leo before this Pope

Making N64 Zelda relativistically correct and calling it Ocarina of Spacetime

Maintaining open-source software that others rely on actually takes time!!!

I learned yesterday that I am a 2025 Kavli Fellow!
https://www.nasonline.org/news/national-academy-of-sciences-selects-the-2025-kavli-fellows/

Wake up babe, new preprint just dropped

Length dependence of waveform mismatch:
a caveat on waveform accuracy
Mitman, Stein, et al.
https://arxiv.org/abs/2502.14025

Screenshot of the title/auth list/abstract:

Length dependence of waveform mismatch:
a caveat on waveform accuracy

Keefe Mitman, Leo C. Stein, Michael Boyle, Nils Deppe, Lawrence E. Kidder, Harald P. Pfeiffer, Mark A. Scheel

Abstract. The Simulating eXtreme Spacetimes Collaboration’s code SpEC can now routinely simulate binary black hole mergers undergoing ∼ 25 orbits, with the longest simulations undergoing nearly ∼ 180 orbits. While this sounds impressive, the mismatch between the highest resolutions for this long simulation is O(10^−1). Meanwhile, the mismatch between resolutions for the more typical simulations tends to be O(10^−4), despite the resolutions being similar to the long simulations’. In this note, we explain why mismatch alone gives an incomplete picture of code—and waveform—quality, especially in the context of providing waveform templates for LISA and 3G detectors, which require templates with O(10^3) − O(10^5) orbits. We argue that to ready the GW community for the sensitivity of future detectors, numerical relativity groups must be aware of this caveat, and also run future simulations with at least three resolutions to properly assess waveform accuracy.

Fig 2 from the preprint, which shows mismatch (between resolutions) as a function of length of time interval over which the mismatch is computed. There are 4 curves, each going roughly like M ~ t^2. The four curves are in two groups: Two from simulations SXS:BBH:1412, and two from SXS:BBH:1132. Each simulation has two different truncation error comparisons: med vs. low and high vs. med. They both show convergence.

@dpiponi This kind of parallels how people think that there's something different about elementary functions vs special functions... or a function that is only defined in terms of an infinite series, or something else that they don't count as "closed form." How do you think your computer finds exp(4.3)? Basically the same approach as any special function e.g. Bessel J_5(4.3).

@johncarlosbaez hah! That's why it was in my subconscious... You linked it and mentioned the title in your post!

@johncarlosbaez can you remind me — did we go over the Wallis-style product for the lemniscate constant? I don't know how it would be derived, and why it's similar to the Wallis product for pi, and if there's a similar product for all the generalizations.

Y'all it's a 🎉 New paper day 🎉 for me, and I wrote an explainer over on bluesky... I don't have the energy to copy it over here. Too many platforms! Check out my explainer at https://bsky.app/profile/duetosymmetry.com/post/3ld2p73zkzs2o or just read the paper at https://arxiv.org/abs/2412.06887.

Screenshot of title/author list/abstract of the paper linked in post

$Figure 1 and caption from the paper. The horz axis is retarded time u after the peak, in units of M. The vertical axis is R|\dot{h}_{20}|/\nu, on a log scale. We see 4 curves, almost exactly on top of each other. Three are from NR simulations at q=1, 2, 3. The fourth is form perturbation theory. They all show an exponentially decaying ringing, followed by a power-law tail. An inset shows the "instantaneous power law index" for each curve, as a function of retarded time, all in rough agreement between -1.5 and -2.$

Figure 2 from the paper, a schematic spacetime diagram. There's a plunging trajectory, outgoing null rays, the radius of an "extraction sphere," and an incoming null ray showing "causal contact with boundary."

A schematic of a spacetime diagram with a green line at 45°, labeled "prompt". From the main right-going "prompt" line, there is a blue left-going scattered wave. From the blue one, there is a red right-going scattered wave. From the red one, an orange left-going wave. From the orange, a yellow right-going wave. You should be able to fill in the gaps and see that these would fill in the whole interior of the light cone.

Chemists/biologists/other folks who use preprint servers besides the arXiv: Do you get bibtex entries with the eprint=... field pointing to servers different from the arXiv?

@jsdodge mathematica 🙈

@uep@octodon.social definitely, those points are the same spacetime interval away from the origin as the top green points. I just didn't draw it to keep the figure simpler. The metric on each "half" is the same, d-dimensional hyperbolic space.

Sometimes I make neat figures: (explanation in alt text)

$A figure of two different types of hyperboloids and a cone, nested in a "natural" way as defined by the geometry of Minkowski space. The surfaces are level sets of the invariant Lorentz interval being constant. When x_\mu x^\mu is negative we get a hyperboloid of two sheets (only the top one is shown, green). When x_\mu x^\mu = 0, we get the light cone of the origin (orange). When x_\mu x^\mu is positive, we get a hyperboloid of one sheet (blue). The text below the figure reads: Figure 2.3: Level sets of the invariant Lorentz interval being timelike (green, top), lightlike (orange), or spacelike (blue, outermost). A quadrant has been cut out for visibility. If the ambient space is (d + 1)-dimensional Minkowski, the geometries on the submanifolds are H^d (green, top), a “null” geometry (orange), and d-dimensional de Sitter spacetime (blue, outermost).$

Leo C. Stein boosted:

Hello nerds (affectionate). I’m working on a project for school that I’d like to program some RFID chips for. Does anybody have a USB RFID Reader that I might be able to borrow for a couple of weeks? If you're in the Bay Area I can come get it in person. If not, I can pay for shipping to and from me.

Client Info