Deyan Ginev's Avatar

Deyan Ginev

@dginev.bsky.social

Software Engineer at arXiv. LaTeXML dev. The Web was created to share science. Let's keep it that way. creator of http://ar5iv.org

363 Followers  |  405 Following  |  304 Posts  |  Joined: 21.08.2023
Posts Following

Posts by Deyan Ginev (@dginev.bsky.social)

Typst Meetup 2026: Keynote
YouTube video by Typst Typst Meetup 2026: Keynote

The first talk of our meetup is live now: In the keynote, we tell-all about our journey from scrappy upstarts to a professional open source project.

New uploads every weekday (Mo-Fri) at 17:00 UTC on YouTube. Subscribe so you don't miss the 10 talks!

www.youtube.com/watch?v=WOQX...

06.03.2026 18:05 β€” πŸ‘ 8    πŸ” 3    πŸ’¬ 1    πŸ“Œ 0

πŸ—“οΈ The February 2026 arXiv articles are now in ar5iv.

06.03.2026 12:27 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

If you really want people to read what you wrote, you will give it away for free, like a Jehova Witness that is handing out bibles.

04.03.2026 19:34 β€” πŸ‘ 7    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0
Preview
GitHub - lab-cosmo/upet: Universal interatomic potentials for advanced materials modeling Universal interatomic potentials for advanced materials modeling - lab-cosmo/upet

Now, this is just a milestone on that path, but it's already something worth sharing, so thanks to arXiv you can read about it arxiv.org/html/2603.02..., and thanks to uPET github.com/lab-cosmo/up... and metatomic, you can try already a universal MLIP trained on it.

03.03.2026 11:15 β€” πŸ‘ 2    πŸ” 2    πŸ’¬ 1    πŸ“Œ 0
Net and Prune: A Linear Time Algorithm for Euclidean Distance Problems1footnote 11footnote 1Work on this paper was partially supported by NSF AF awards CCF-0915984 and CCF-1217462. The full updated ve...

A while back, we realized that one can (1+Ξ΅)-approximate the Hausdorff distance between point sets in ℝ^d, in linear time. Unfortunately, this was already done implicitly in a remark in a published old paper. So, we updated the arXiv version:
arxiv.org/html/1409.74...

03.03.2026 08:23 β€” πŸ‘ 4    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0

13 years ago, Aaron Swartz came to visit the Internet Archive during his case. I had time with him, got a picture with him, told him that if they were stupid enough to send him to candy jail he'd become a folk hero and he could write a book that would change the world. He said he appreciated it.

03.03.2026 01:59 β€” πŸ‘ 147    πŸ” 30    πŸ’¬ 4    πŸ“Œ 0

I need an AGENT.md file automating step 3

02.03.2026 19:48 β€” πŸ‘ 6    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

I see the shortcut, but where's the final monetization coming from?

02.03.2026 19:38 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0
The arXiv of the future will not look like the arXiv

🌷Spring cleaning for arXiv's HTML Papers

1️⃣ The header on new articles is now collapsible and a bit more compact on mobile.

2️⃣ We've also moved the warnings and caveats out of the way, in favor of focused reading.

Let us know what you think? #arXiv

arxiv.org/html/1709.07...

02.03.2026 15:29 β€” πŸ‘ 3    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0
Preview
TeX Live 2026 released | There and back again Get the Champagne ready, we have released the final images of TeX Live 2026. Nothing spectacularly new in this version, but improvements across that bank, and a new player: `xdvipsk`, an extended dvip...

TeX Live 2026 is released - mirror updates are running! #TeXlive www.preining.info/blog/2026/03...

01.03.2026 23:08 β€” πŸ‘ 5    πŸ” 4    πŸ’¬ 0    πŸ“Œ 0
A picture of Joe Halpern smiling in green shirt in front of a blue background.

A picture of Joe Halpern smiling in green shirt in front of a blue background.

Today arXiv remembers our colleague Joe Halpern, who was instrumental in founding arXiv's CS section.

Joe's passions ranged far & wide and we're lucky that arXiv was one of them. Joe, thank you for giving so much to arXiv - you are missed.

blog.arxiv.org/2026/02/27/remembering-joe-halpern

27.02.2026 18:38 β€” πŸ‘ 50    πŸ” 12    πŸ’¬ 2    πŸ“Œ 2

If you find pieces of open infrastructure valuable - especially fledgling ones - remember to voice your support in forums that matter.

And send kudos to their maintainers when you can. A couple of encouraging words can make a real difference.

28.02.2026 15:21 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

We're looking for a Senior Software Developer (frontend) to join Crossref.

Remote, full-time.

Deadline: March 10, 2026. https://www.crossref.org/jobs/2026-02-17-senior-software-developer-frontend

#jobs #hiring #remote #remotework

28.02.2026 01:34 β€” πŸ‘ 3    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0
Attention Authors: Updates for .bib file processing and TeX in arXiv submissions – arXiv blog

Swinging by to happily note upgrades landed in both tracks, as everyone was hoping. Here's the Nov blog post that announced xelatex support:

blog.arxiv.org/2025/11/05/a...

28.02.2026 01:07 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0
Preview
Nanoscience is latest discipline to embrace large-scale replication efforts A European project calls for help to verify whether carbon quantum dots are really able to sense chemicals in cells.

Wonderful to see this replication effort in the physical sciences using the models of many labs, preregistration, and transparency that have benefitted other fields.

And, an investment of $9.5 million to do it!

www.nature.com/articles/d41...

22.02.2026 13:45 β€” πŸ‘ 37    πŸ” 11    πŸ’¬ 0    πŸ“Œ 0
Preview
No shortcuts to research information citizenship - Digital Science Being open isn't enough - true "research information citizenship" requires a robust, genuinely open research infrastructure.

"For open infrastructure, the burden of AI harvesting is existential...a swarm of AI scrapers mirror entire collections, the cost of being open rises sharply...degraded performance for legitimate users, polluted metrics, staff time diverted to firefighting" www.digital-science.com/blog/2026/02...

22.02.2026 02:36 β€” πŸ‘ 14    πŸ” 8    πŸ’¬ 0    πŸ“Œ 1

Eminem's still faster.

20.02.2026 02:04 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Especially science. It's so sad to be publishing so many papers that you're too busy to do science.

20.02.2026 01:20 β€” πŸ‘ 29    πŸ” 4    πŸ’¬ 0    πŸ“Œ 0
19.02.2026 19:43 β€” πŸ‘ 72    πŸ” 11    πŸ’¬ 0    πŸ“Œ 0
Cornell University, Computer Science Job #AJO31698, Lecturer/Senior Lecturer, Computer Science, Cornell University, New York, New York, US

Care about preparing people to contribute responsibly to building the next generation of AI and technology?

Full-time (or at least 50%) lecturer position at Cornell Tech just posted, teaching computer science or related topics.

academicjobsonline.org/ajo/jobs/31698

17.02.2026 22:37 β€” πŸ‘ 12    πŸ” 10    πŸ’¬ 0    πŸ“Œ 2
We want to evaluate
$$
\sum_{\color{red}k=0}^\infty (\color{red}k+1) \color{blue}p^{\color{red}k}\,.
$$
Introduce the function $f$, for $|\color{blue}x|<1$:
$$
f(\color{blue}x) = \sum_{\color{red}k=0}^\infty \color{blue}x^{\color{red}k}\,.
$$
That's a nice geometric series, and we easily get $f(\color{blue}x) = \frac{1}{1-\color{blue}x}$. So we can differentiate that:
$$
f'(\color{blue}x) = \frac{1}{(1-\color{blue}x)^2} 
$$
But $f$ was defined as a power series, and we can also differentiate *that* termwise:
$$
f'(\color{blue}x) = \sum_{\color{red}k=1}^\infty \color{red}k \color{blue}x^{\color{red}{k-1}} = \sum_{\color{red}k=0}^\infty {(\color{red}k+1)} \color{blue}x^{\color{red}{k}}\,.
$$
Well, $f'(\color{blue}x)= f'(\color{blue}x)$ (!), so we can use both expressions, and evaluate them at $\color{blue}p$:
$$
\boxed{\sum_{\color{red}k=0}^\infty {(\color{red}k+1)} \color{blue}p^{\color{red}{k}}
= \frac{1}{(1-\color{blue}p)^2}}
$$

We want to evaluate $$ \sum_{\color{red}k=0}^\infty (\color{red}k+1) \color{blue}p^{\color{red}k}\,. $$ Introduce the function $f$, for $|\color{blue}x|<1$: $$ f(\color{blue}x) = \sum_{\color{red}k=0}^\infty \color{blue}x^{\color{red}k}\,. $$ That's a nice geometric series, and we easily get $f(\color{blue}x) = \frac{1}{1-\color{blue}x}$. So we can differentiate that: $$ f'(\color{blue}x) = \frac{1}{(1-\color{blue}x)^2} $$ But $f$ was defined as a power series, and we can also differentiate *that* termwise: $$ f'(\color{blue}x) = \sum_{\color{red}k=1}^\infty \color{red}k \color{blue}x^{\color{red}{k-1}} = \sum_{\color{red}k=0}^\infty {(\color{red}k+1)} \color{blue}x^{\color{red}{k}}\,. $$ Well, $f'(\color{blue}x)= f'(\color{blue}x)$ (!), so we can use both expressions, and evaluate them at $\color{blue}p$: $$ \boxed{\sum_{\color{red}k=0}^\infty {(\color{red}k+1)} \color{blue}p^{\color{red}{k}} = \frac{1}{(1-\color{blue}p)^2}} $$

Let's say you want, e.g., to compute the expectation of a Geometric r.v. That'll involve, at some point, evaluating a series of the form "Ξ£ (k+1) p^k" which looks like what Lovecraft may have done to a geometric series. How to do it?

One trick I enjoy: differentiate the same function, in two ways!

18.02.2026 12:27 β€” πŸ‘ 38    πŸ” 6    πŸ’¬ 1    πŸ“Œ 0

any sufficiently advanced technology is indistinguishable from language

16.02.2026 02:00 β€” πŸ‘ 56    πŸ” 11    πŸ’¬ 3    πŸ“Œ 1

Exactly right and apologies!
The latest LaTeXML has a precompiled kernel to avoid this pain.

ar5ivist uses it (and tracks arXiv's output), so you can either run its Dockerfile directly or reuse the install steps. Down to 1-2 seconds for loading expl3:

github.com/dginev/ar5iv...

10.02.2026 12:22 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

πŸ—“οΈ The January 2026 arXiv articles are now in ar5iv.

06.02.2026 12:51 β€” πŸ‘ 2    πŸ” 1    πŸ’¬ 0    πŸ“Œ 1

πŸ€―β€œI tasked 16 agents with writing a Rust-based C compiler, from scratch, capable of compiling the Linux kernel. Over nearly 2,000 Claude Code sessions and $20,000 in API costs, the agent team produced a 100,000-line compiler that can build Linux 6.9 on x86, ARM, and RISC-V.”

05.02.2026 22:56 β€” πŸ‘ 17    πŸ” 3    πŸ’¬ 5    πŸ“Œ 1
Preview
Maintaining the Bridges Thoughts and analogies about infrastructure...

Hot off the blog presses: Maintaining the Bridges

Thoughts and analogies about infrastructure...

bkardell.com/blog/Bridges...

02.02.2026 18:19 β€” πŸ‘ 5    πŸ” 3    πŸ’¬ 0    πŸ“Œ 0

yesterday i gave my first talk at fosdem on mathml interoperability! i was quite nervous and there were internet issues, but it was a nice experience :3

you can watch it here: https://video.fosdem.org/2026/h1309/NJM3KB-mathml-core.av1.webm :blobhaj_reach:

01.02.2026 06:57 β€” πŸ‘ 9    πŸ” 5    πŸ’¬ 2    πŸ“Œ 0
Video thumbnail

❀️

30.01.2026 19:57 β€” πŸ‘ 28069    πŸ” 9224    πŸ’¬ 362    πŸ“Œ 760
Preview
OpenAI’s latest product lets you vibe code science Prism is a ChatGPT-powered text editor that automates much of the work involved in writing scientific papers.

β€œThe idea is to put ChatGPT front and center inside software that scientists use to write up their work in much the same way that chatbots are now embedded into popular programming editors.

It’s vibe coding, but for science.”

27.01.2026 21:52 β€” πŸ‘ 245    πŸ” 39    πŸ’¬ 43    πŸ“Œ 64

Of all the strange second-order effects of co-pilot LLMs, saving TeX was the one that shocked me the most.

28.01.2026 12:13 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 1