iiinova — where your digital self keeps evolving

Daan van Esch @daanvanesch.nl · 03/10/2026

It was fun chatting at Interspeech, enjoy the Blue Mountains! If you happen to find yourself in a bookstore there:

A copy of the book The Cats of Australia by Jodie Stewart

Daan van Esch @daanvanesch.nl · 29/07/2026

Hard to beat the recommendation above re: relativity theory for babies (I've read it and found it was just about at my level), but I also really enjoyed Tobias Hürter, Das Zeitalter der Unschärfe.

110

Daan van Esch @daanvanesch.nl · 20/07/2026

I lived in California 2015-2020. The first time I ever encountered chicken and waffles on a menu was on April 1, as a daily special. I turned to some coworkers and said, haha, that's a good joke but I'm not falling for that. It took some convincing before I accepted that it was a normal dish...

011

Daan van Esch @daanvanesch.nl · 03/07/2026

Ze zeggen nu net minstens tot zondagavond 🙃

110

Daan van Esch @daanvanesch.nl · 03/07/2026

Yet no one deserves it more! Looking forward to getting to congratulate you in person later today!!

110

Reposted by Daan van Esch

Julia Kreutzer @juliakreutzer.bsky.social · 30/06/2026

🤨"But why linguistics" is the most common question when talking about linguistic reasoning benchmarks. Last year we organized a shared task at WMT...and no one participated 🤣 🤯Let me change your mind why this is one of the most challenging, focused and best reasoning benchmarks right now.

1142

Reposted by Daan van Esch

Julia Kreutzer @juliakreutzer.bsky.social · 30/06/2026

🔥Possibly the most fun and underrated AI challenge this summer: Compete on unseen linguistic reasoning problems and present your solutions to the expert jury in a month!

011

Reposted by Daan van Esch

Google for Developers @developers.google.com · 09/06/2026

Gemini 3.5 Live Translate is now available in public preview on the Gemini API and Google AI Studio. 💬 This model translates speech as it streams, giving developers a blazing-fast, low-latency engine to build some seriously cool audio apps. See it in action 👇

7349

Daan van Esch @daanvanesch.nl · 09/06/2026

(Ik stuur jullie even een mailtje, ik heb de hele tekst van al die artikelen in Feedly)

000

Daan van Esch @daanvanesch.nl · 09/06/2026

Ik heb ze zo te zien nog allemaal in mijn Feedly RSS cache, zal ik jullie een zipje sturen ofzo?

100

Daan van Esch @daanvanesch.nl · 28/04/2026

Cool, I'll definitely check it out! Also tagging @very-laurie.bsky.social who will likely enjoy this

020

Daan van Esch @daanvanesch.nl · 18/04/2026

I've always wondered if we wouldn't conclude that most (all?) of the world's languages were polysynthetic if we focused exclusively on analyzing spoken language data -- that is, is polysynthetic even well-defined if we forget about the existence of orthographies? What even is a word without writing?

110

Reposted by Daan van Esch

Julia Kreutzer @juliakreutzer.bsky.social · 03/03/2026

💭We need more research that focuses on aspects beyond accuracy, especially in multilingual AI. 👉Help us explore the importance of culture in building and testing AI, with a few minutes of your time. Happy to have a chat as well with anyone who's interested in that space!

163

Daan van Esch @daanvanesch.nl · 27/02/2026

Mooie herinneringen aan mijn bezoek een paar jaar geleden!

010

Daan van Esch @daanvanesch.nl · 22/02/2026

Okay AfroLID is working now! It's also a quantized model, but it's still 200MB so it's a bit hefty for a metered connection...maybe I'll put it behind a button you can click to load it explicitly. The FastText models go down to ~15MB each so that still feels acceptable without explicit UI action

010

Daan van Esch @daanvanesch.nl · 22/02/2026

...know if you have any suggestions / feature requests / etc, or if there are models that I should be adding. BTW, caveat in case it's not obvious yet from the page (I'll also make this clearer), the FastText models are quantized to fit in-browser so they won't deliver the exact same results

110

Daan van Esch @daanvanesch.nl · 22/02/2026

Thanks for spreading the word! Inspired by your great paper :) I still need to fix AfroLID inference and I also want to spend a bit of time improving the intro at the top of the page, adding references to all the LID models, and so on. So definitely work in progress, but don't hesitate to let me...

110

Reposted by Daan van Esch

Daniel van Strien @danielvanstrien.bsky.social · 19/02/2026

Re-OCR'd the complete 1771 Encyclopaedia Britannica (2,724 pages) with a single command on @hf.co Jobs. - 0.9B model (GLM-OCR) ~$0.002/page ~$5 total on an L4 GPU Before (old Tesseract ocr) → After

Screenshot of old vs new ocr.

old ocr text is garbled. New ocr much cleaner.

59616

Reposted by Daan van Esch

Julia Kreutzer @juliakreutzer.bsky.social · 18/02/2026

🌱Very proud of our team's latest release 😊 meet Tiny Aya, a massively multilingual model with 3.35B parameters. Tech report here: github.com/Cohere-Labs/...

github.com

1327

Daan van Esch @daanvanesch.nl · 17/02/2026

It's just a static HTML file with all the CSS and JS embedded, and it runs in-browser, so it's easy to move around, and it should be reasonably straightforward to have multiple copies pointing to multiple preconfigured EAF files (plus audio) so you can like, potentially show a corpus or something.

110

Daan van Esch @daanvanesch.nl · 17/02/2026

daanvanesch.nl/eaf-viewer.h... is the one I was playing with the other day, it currently scrolls horizontally but I'm sure Gemini/Codex/Claude would quickly make that vertical instead. Haven't had the chance to test it on a lot of EAF files yet so it may not work out-of-the-box but happy to help!

daanvanesch.nl

EAF Viewer

110

Daan van Esch @daanvanesch.nl · 17/02/2026

Cool yeah that should work, let me dig it up! BTW there's also brownclps.github.io/LingView/#/s... which is github.com/BrownCLPS/Li... -- definitely also an option especially if you're comfortable poking around in the terminal, see github.com/BrownCLPS/Li...

brownclps.github.io

LingView

110

Daan van Esch @daanvanesch.nl · 17/02/2026

Two great groups teaming up, looking forward to seeing the impact you'll be delivering together!

020

Daan van Esch @daanvanesch.nl · 17/02/2026

I (had my coding agent) put something together the other day that did this for some ELAN eaf files, is that what you're using as well? I can dig it up, probably quite doable for like, a Praat TextGrid file too

110

Daan van Esch @daanvanesch.nl · 17/02/2026

I'm not sure about WordPress plugins etc but for a format like ELAN eaf files it's reasonably straightforward to have one of the modern coding agents whip something up if you're just working without a CMS like WordPress in the loop. Is it an existing page you'd want to add the time-aligned text to?

120

Daan van Esch @daanvanesch.nl · 13/02/2026

Great to see this amazing collaborative work on an absolutely key problem in building tech that works well across the world's languages: language classification in web text. Often ignored, it's still one of my personal favorite areas to work in. Congrats and thank you to everyone who worked on this!

021

Reposted by Daan van Esch

eleutherai.bsky.social @eleutherai.bsky.social · 13/02/2026

Why care about LangID on crawled data? It's the first gate in the multilingual data pipeline. If your LID model misclassifies a low-resource language as noise or confuses it with a related high-resource one, that language doesn't make it into your corpus. Bad LangID = no data.

161

Daan van Esch @daanvanesch.nl · 13/02/2026

Oh I see, a lot of them are controlled manually, got it! Still, always a nice city to visit so maybe I'll squeeze it in somewhere in the next few weeks. Thanks!

010

Daan van Esch @daanvanesch.nl · 13/02/2026

That's cool! I've always wanted to learn more about premodern automata like this, supposedly there were also some mechanical marvels in the Tang dynasty. Looks like I'll have to make my way over to Aachen sometime soon!

110

Daan van Esch @daanvanesch.nl · 09/02/2026

But when I was in Leuven last year and asked (in an otherwise Dutch sentence) for a pain au chocolat at the bakery, the lady did laugh and say that there's no need to pull out fancy French words. To me it's very standard in nl-nl to just use the French name. Now I know the proper Flemish Dutch name!

240

Daan van Esch @daanvanesch.nl · 09/02/2026

In my Netherlandic Dutch I always just call this a "pain au chocolat" and that's also what some brands of baked goods at AH call them, but the ones AH sells from their own bakery are "chocoladebroodjes". I'd never heard "chocoladekoek".

250

Reposted by Daan van Esch

Interspeech 2026 @interspeech.bsky.social · 05/02/2026

What are roles of speech science and technology projects in advancing Indigenous cultural vitality and self-determination? Participate in this important discussion in the Special Session 'Indigenous Voices in Speech Sciences and Technology' at #Interspeech2026 indigenousvoicesinterspeech.github.io

024

Reposted by Daan van Esch

Sung Kim @sungkim.bsky.social · 26/01/2026

The intelligence part suddenly feels quite a bit ahead of all the rest of it - integrations (tools, knowledge), the necessity for new organizational workflows, processes, diffusion more generally. 2026 is going to be a high energy year as the industry metabolizes the new capability."

041

Reposted by Daan van Esch

Abdoulaye Diack @diack.bsky.social · 16/01/2026

Meet TranslateGemma. 💎 ✅ Open weights (4B, 12B, 27B) ✅ 55 languages + 100s more in training data ✅ Multimodal capabilities (image text) Blog: blog.google/innovation-a... Paper: arxiv.org/pdf/2601.09012 Model: huggingface.co/collections/... Cookbook: colab.research.google.com/github/googl...

blog.google

TranslateGemma: A new suite of open translation models

TranslateGemma is a new family of open translation models built on Gemma 3.

1384

Daan van Esch @daanvanesch.nl · 11/01/2026

...wherever we can (so you can see if we used e.g. Glottolog's lat-long or some other source) but agreed that even more detailed descriptions of how each speaker estimate and writing system was arrived at would've been great. Still we hoped that on the whole it might be useful, but totally agree!

110

Daan van Esch @daanvanesch.nl · 11/01/2026

it'd be something we'd release publicly. This was part of the work we did for doi.org/10.48550/arX... and then we eventually ended up releasing it publicly in 2022 and 2024 LREC papers. In the detailed LinguaMeta JSON files for each language we do provide a back reference...

doi.org

Writing Across the World's Languages: Deep Internationalization for Gboard, the Google Keyboard

This technical report describes our deep internationalization program for Gboard, the Google Keyboard. Today, Gboard supports 900+ language varieties across 70+ writing systems, and this report descri...

120

Daan van Esch @daanvanesch.nl · 11/01/2026

...but it's quite challenging, also involving lots of non-academic sources, and sometimes it's also just a judgment call. I wish we'd documented it in a more structured way looking back! But when we were working on this like 10 years ago it didn't occur to me at the time that someday...

doi.org

Writing Across the World's Languages: Deep Internationalization for Gboard, the Google Keyboard

110

Daan van Esch @daanvanesch.nl · 11/01/2026

Yeah, totally agree, this is one of those things that retroactively I wish I'd done...the back story is that when we were working to figure out what languages to add to Gboard (Google's Android keyboard; hence the writing systems), we pored over countless resources to create best-effort estimates...

110

Daan van Esch @daanvanesch.nl · 11/01/2026

Our goal is definitely not to replace Glottolog or any of the existing amazing resources, we just thought it'd be good to have a public version of our internal visualization that we use to give a broader (non-linguistics) audience a sense of the world's rich linguistic diversity. Happy to chat more!

110

Daan van Esch @daanvanesch.nl · 11/01/2026

Let me check on this on Monday, would be great to make the table (and paper) Lorena linked to more easily accessible from the visualization site. We're huge Glottolog fans and the visualization was intended to live alongside the paper (presented at LREC 2024) but we'll make the link clearer!

110

Daan van Esch @daanvanesch.nl · 09/12/2025

In Aosta wordt historisch ook Arpitaans gesproken geloof ik dus wellicht heeft het daarmee te maken? In ongeveer dezelfde hoek van de wereld, in Monaco, bordjes in het Frans en het Monegaskisch (een soort Ligurijns, vroeger tot in Nice gesproken):

130

Daan van Esch @daanvanesch.nl · 09/12/2025

In het gebied in de Blue Ridge Mountains waar Cherokee wordt gesproken:

110

Reposted by Daan van Esch

Simon Willison @simonwillison.net · 28/11/2025

Couldn't have built them without the AI assistance? Yes, if I had unlimited time - but I don't have unlimited time, so I'm happy to settle with being able to read through, understand and explain what they did in my behalf instead

1261

Reposted by Daan van Esch

Simon Willison @simonwillison.net · 28/11/2025

AI assistance entirely changes that equation - if I can reduce a problem to something that a coding agent can go and crunch away at while I'm doing other things I can say "yes" to all manner of learning exercises that I previously didn't have enough time to take on

1172

Daan van Esch @daanvanesch.nl · 22/11/2025

Your books are also having a great time chatting with each other in the linguistics section here at the Kinokuniya bookstore on Orchard Road in Singapore!

1130

Daan van Esch @daanvanesch.nl · 21/10/2025

Ik doe tegenwoordig precies hetzelfde: ik dicteer in mijn coding agent. En dat gaat lastiger op kantoor dan thuis op mijn werkkamer. Voor mij is er zowat genoeg verschil in mijn productiviteit ten opzichte van typen om een dagje meer thuis te werken...

120

Reposted by Daan van Esch

Verena Blaschke @verenablaschke.bsky.social · 21/10/2025

VarDial 2026 will be colocated with @eaclmeeting.bsky.social! We're looking forward to your papers on NLP for similar languages, varieties and dialects :) Deadline: Dec 19 (Jan 2 for pre-reviewed ARR papers) sites.google.com/view/vardial...

VarDial @ EACL 2026, with important dates (see next post for text version).
Photo CC-0.

11410

Daan van Esch @daanvanesch.nl · 20/10/2025

I've been watching Feuer und Flamme season 10 in Heidelberg the last few weeks, which captures the everyday work of the local fire service. Definitely needs dialectal ASR for some speakers!

020

Reposted by Daan van Esch

Morris Alper @malper.bsky.social · 11/10/2025

The ConlangCrafter pipeline harnesses an LLM to generate a description of a constructed language and self refines it in the process. We decompose language creation into phonology, grammar, and lexicon, and then translate sentences while constructing new needed grammar points.

182