iiinova — where your digital self keeps evolving

DLD @davidlowryduda.bsky.social · 16/09/2026

I've somehow found myself with some devops responsibilities. Today is rotate-all-the-keys day. It's not hard, but it is annoying.

000

DLD @davidlowryduda.bsky.social · 16/09/2026

I've found that this type of OCR problem (where the text from the middle column is ignored and reading order goes from the first column to the third, as in your example) to be challenging to overcome consistently.

000

DLD @davidlowryduda.bsky.social · 11/09/2026

But can AI ever give me the same satisfaction as watching "tail -f my_great_process.log" on a screen?

001

Reposted by DLD

Ted Underwood @tedunderwood.com · 10/09/2026

Heads-up everyone in digital humanities and/or cultural AI: this is a hugely valuable resource. It does the systematic cleaning and segmentation — of a million books — that we have always wanted to have, for computational research, and have never quite produced in a way that stuck.

18027

DLD @davidlowryduda.bsky.social · 10/09/2026

I see this in the library space too. Many companies want the data and to plant flags without devoting effort to preservation or provenance or sustainability more broadly. How can we move towards a more effective and beneficial inclusion of LLMs and other AI tools? 3/3

000

DLD @davidlowryduda.bsky.social · 10/09/2026

But I don't think that fast and am still formulating my thoughts. I know that I do not agree with the anti-AI positions I read. On the other hand, I feel that some recent LLM efforts are non-sustainable and ultimately bad for math and science and art and the humanities. 2/3

110

DLD @davidlowryduda.bsky.social · 10/09/2026

Navier-Stokes, Alpoge-Buckmaster, and OpenAI have made bigger waves than I expected. I've read many takes and opinions. 1/3

100

DLD @davidlowryduda.bsky.social · 09/09/2026

Should I take the fact that none of my work or research has been scooped as a sign that it's not exciting enough?

000

DLD @davidlowryduda.bsky.social · 08/09/2026

I recently heard someone talk about how math has long straddled the humanities and other sciences. Then the atomic bomb made govs throw money at math and physics, and mathematicians applied for these pots of money. And now mathematicians are saying it's the human part that's valuable (it is!).

020

DLD @davidlowryduda.bsky.social · 08/09/2026

At 0% of their word: no scientist hoping to get credit can use OpenAI (or maybe other LLM providers) without fear of being scooped. At 100% of their word: no scientist can work in the open without fear of being scooped. Both suppress communication.

1223

Reposted by DLD

IDI @institutional.org · 08/09/2026

Institutional Books is expanding. The Institutional Data Initiative @institutional.org at @harvardlawadmin.bsky.social Library has added paragraph-level text analysis and 22M extracted images to our work on ∼1M books, opening new possibilities for research, experimentation, and discovery. 🧵

1288

DLD @davidlowryduda.bsky.social · 08/09/2026

I previously read the first few paragraphs and thought not much of it. That's because all the content starts on the second page

020

Reposted by DLD

Clément Canonne @ccanonne.github.io · 08/09/2026

See cims.nyu.edu/~tristanb/st... for what one side is saying, esp. page 2 onwards.

cims.nyu.edu

1162

DLD @davidlowryduda.bsky.social · 08/09/2026

I hear fears about the interaction of LLMs and the sciences. People worry how the sciences will change. One challenge with discussing this is that it's hard to describe *how the sciences work now.* #mathematics #science

000

DLD @davidlowryduda.bsky.social · 04/09/2026

A few years ago, I helped formalize a very small subset of FLT (as an introduction to lean). My work consisted of perhaps 20 or so hours of actual work over a few days. I felt that doing the full FLT would be very doable, but tedious and not enjoyable.

010

DLD @davidlowryduda.bsky.social · 04/09/2026

Fermat's Last Theorem was formalized by Anthropic www.anthropic.com/research/for... A mostly-automated, distributed swarm of LLM agents worked on this for 11 days.

anthropic.com

Formalizing Fermat's Last Theorem

Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.

110

DLD @davidlowryduda.bsky.social · 08/09/2025

Can one then show that 1 = 2, say by manipulating a - (a + 1) = 0 into a = a + 1, adding 1 to both sides, and subtracting a? No! The reason why is that LEAN knows that subtraction in the naturals isn't subtractive. (Or rather it doesn't think that it is subtractive). (3/3)

110

DLD @davidlowryduda.bsky.social · 08/09/2025

The proof in LEAN is very simple: tell it to use 0. What happens behind the scenes is that, in LEAN, 0 - 1 = 0 (when considered in natural numbers). I think this is because it's convenient to have naturals closed under basic operations. (2/3)

100

DLD @davidlowryduda.bsky.social · 08/09/2025

At this morning's opening session of the code4math opening (preview.scholarlattice.org/collections/...), I learned that in LEAN one can prove the following theorem: There exists a natural number a such that a - (a + 1) = 0. (1/3)

preview.scholarlattice.org

131

DLD @davidlowryduda.bsky.social · 16/04/2025

+1 for Secret Mall Apartment

000

DLD @davidlowryduda.bsky.social · 20/12/2024

It definitely contributes to math-as-mysticism, incantations done by all-knowing mages that is intractable to mere mortals. But this is far from the truth! And we want more people to recognize that mathematical thinking is useful and approachable and worthwhile!

031

DLD @davidlowryduda.bsky.social · 18/12/2024

And also this was the source of bsky.app/profile/davi...

000

DLD @davidlowryduda.bsky.social · 18/12/2024

I just submitted "The Fibonacci Zeta Function and Continuation" (or as I think of it it, Fibonacci Zeta Function I) to the arxiv. Cheers to my collaborators Eran Assaf, Chan Kuan, and Alex Walker. I feel like I've accomplished something, so I will goof off with my daughter for a while.

210

Reposted by DLD

Leo C. Stein @duetosymmetry.com · 14/12/2024

Pro tip for folks who use beamer: Turn off the nav symbols! \setbeamertemplate{navigation symbols}{} 🧪⚛️🧮

1255

Reposted by DLD

code4math Community @code4math.org · 18/11/2024

Hey #MathSky! 👋 Interested in learning how to leverage #computing to advance your mathematics research and teaching? You may be interested in the @aimathematics.bsky.social sponsored "Leveraging GitHub and AI for Mathematics Research and Teaching" PEP: jointmathematicsmeetings.org/meetings/nat...

jointmathematicsmeetings.org

Join Us at the Joint Mathematics Meetings - The Largest Mathematics Gathering Globally

Discover cutting-edge mathematical advancements and network with industry leaders at the world's largest mathematics meeting!

275

DLD @davidlowryduda.bsky.social · 07/12/2024

Thanks. I've never heard of Ologs before.

010

DLD @davidlowryduda.bsky.social · 06/12/2024

I'm very interested in the different notetaking systems that mathematicians, scientists, programmers, and engineers use. This is talked about somewhere, right? I'm looking for more than "I use a notebook" or "I use [app name]". I want to know how people actually use and organize their notes.

210

DLD @davidlowryduda.bsky.social · 04/12/2024

This is the Fibonacci zeta function. Actually, it's the odd-indexed Fibonacci zeta function, but that's ok. davidlowryduda.com/odd-fibonacci/

A plot of the Fibonacci zeta function. It looks a bit like dunes in a sandy desert, with regular lumps corresponding to poles.

120

DLD @davidlowryduda.bsky.social · 08/03/2024

I appreciate the tip, thanks!

010