j soma (@dangerscarf) 's Twitter Profile
j soma

@dangerscarf

data head, python kid, prof @columbiajourn, director @ledeprog, co-founder @bkbrains + @catrepublicbk. collecting all cats, widening all gyres

ID: 15921550

linkhttps://beacons.ai/jsoma calendar_today20-08-2008 19:00:32

6,6K Tweet

3,3K Followers

2,2K Following

hinterlander (@yoltartar) 's Twitter Profile Photo

this is why i think no llm can ever be skilled in the physical world: the whole technology is "based on a (based on a (based on a true story) story) story"

this is why i think no llm can ever be skilled in the physical world: the whole technology is "based on a (based on a (based on a true story) story) story"
Global Investigative Journalism Network (@gijn) 's Twitter Profile Photo

Summer Investigative Reporting Course Columbia Journalism School will offer 3 scholarships in NYC this July, funded by TX Group. Open to investigative reporters and editors living & working in Asia, Africa, and Latin America. Apply by April 30: twp.ai/4imEgh

Summer Investigative Reporting Course

<a href="/columbiajourn/">Columbia Journalism School</a> will offer 3 scholarships in NYC this July, funded by TX Group. Open to investigative reporters and editors living &amp; working in Asia, Africa, and Latin America.

Apply by April 30: twp.ai/4imEgh
Jireh Deng 鄧以樂 (@bokchoy_baobei) 's Twitter Profile Photo

i'm writing a story for Business Insider about international students who've had their visas revoked or funding cut to their program. if that's you, please email me at [email protected] or shoot me a secure tip via Signal at jdeng.20. 🙏

Vik Paruchuri (@vikparuchuri) 's Twitter Profile Photo

We shipped an alpha version of the new Surya OCR model. No hype, just facts: - 90+ languages (focus on en, romance langs, zh, ar, ja, ko) - LaTeX and formatting - Char/word/line bboxes - ~500M non-embed params - 10-20 pages/s

We shipped an alpha version of the new Surya OCR model.  No hype, just facts:

- 90+ languages (focus on en, romance langs, zh, ar, ja, ko)
- LaTeX and formatting
- Char/word/line bboxes
- ~500M non-embed params
- 10-20 pages/s
j soma (@dangerscarf) 's Twitter Profile Photo

New Natural PDF release, nudging us up to v0.1.10! 📒 Automatically categorize documents with a fun new `.classify` ability 👁️ Differentiate using text on the page *or* vision 🗒️ Works on individual pages or entire collections of PDFs

New Natural PDF release, nudging us up to v0.1.10!

📒 Automatically categorize documents with a fun new `.classify` ability
👁️ Differentiate using text on the page *or* vision
🗒️ Works on individual pages or entire collections of PDFs
j soma (@dangerscarf) 's Twitter Profile Photo

I think today's important updates to Natural PDF are nested .find_all queries and making semantic search 100x more lightweight, buuuut unfortunately the best visuals come from region geometry

I think today's important updates to Natural PDF are nested .find_all queries and making semantic search 100x more lightweight, buuuut unfortunately the best visuals come from region geometry