Charles Loder

link (🤞the new og image works): hebrewtransliteration.app/?test=1

🎉 Version 3 of Hebrew Transliteration App is live! Besides some major tech changes (Astro & Svelte), the UX has been greatly improved. There are far fewer bugs related to how schemas are handled, including improved support for callback functions. Let me know what you think!

Only Firefox supports animated favicons it looks like

Got a redesign cooking, and checkout that favicon!

Want an easy way to find Scripture references in texts? Now you can. I modified our AI pipeline to perform scripture detection. It's fully open-source and found here (next post). It leverages Gemini 3.1 Pro Preview (API key required).

@wjb_mattingly That link didn't work, and I definitely need to see this!

GitHub Repo: github.com/yale-ch/script…

@jonathanahlgren @DrJimHamilton @SBTS Dr. Betts has a goatee now! Also congrats!

I am happy to announce that I have successfully completed and defended my dissertation: Allegiance to David as the Righteous Suffered: The Parallel Characterization of Abigail and Jonathan. Thank you @DrJimHamilton for your instruction and support. Come and study at @SBTS!

While Gemini 3.1 is great, they really dialed up the sycophancy. One of the things I like about Gemini is how much it didn't glaze you

At The Ohio State University, the study of ancient Northwest Semitic languages has moved out of the textbook and into the digital laboratory of epigraphic training. dlato.osu.edu/blog/teaching-…

Dumped a bunch of Khan's writing about the shewa from TTPT into Sonnet 4.6 and has it generate a flowchart. There are errors (partly due to maybe not giving it enough context), but a fun exercise in #digitalhumanities claude.ai/public/artifac…

@Timothy_A_Lee Congratulations!

PhD graduation! 🎓

The barrier to software engineering used to be skill, now it's money

The V[ictoria]&A[albert Museum] has acquired an early watch page of the video sharing platform YouTube and its first-ever video, 'Me at the zoo', uploaded on 23 April 2005. vam.ac.uk/articles/acqui…

@AdlerYonatan I think where it is important when dealing with these tools—ChatGPT, Claude, Gemini, etc.—are essentially meaningless terms. Those aren't models; those are products. The Gemini 3.1 Pro model does a very good job, but Gemini 2.5 may fail.

I just plugged this into 4 different LLMs to test them: ChatGPT Gemini Grok Claude Not one of them understood this pretty straightforward text. It goes without saying that all of them gave answers. Wrong answers, but answers nonetheless...

My daughter referred to New Jersey as "south New York"…and I'm trying to figure out if she's wrong

Home owner pro tip: figure out what you need from Home Depot BEFORE going there. You're welcome

Strunk and White would've have hated this

@bdenckla 🤯 that's awesome

@charles_loder Wouldn't it be nice for the computer to be able to zoom you in on where a word probably is in the Aleppo Codex? I can now do that for any word in the 24 pages of Job, because I painstakingly added some annotations to MAM describing the physical layout of each page. For example:

Here's a screen snip from an Aleppo Codex line break editor I had Claude Opus 4.6 write me in Visual Studio. @charles_loder you'll "get" what this is, I bet!

@bdenckla That's awesome! I'm gonna have to try to see if Gemini Agentic Vision it can create a "digital page" of L or A

@charles_loder I wish! Not as automated as you suggest: I had it write me an editor that allowed *me* to divide the text into the layout found in the AC (two columns since I'm working on Job). So, AI coded the app that helped me use good old human vision (mine) for markup!