Examples of mislabeled web text by existing LangID systems. A full text version is available on the blog post below.
Examples of mislabeled web text by existing LangID systems. A full text version is available on the blog post below.
Language identification still proves to be a challenging task, especially for web data. In collaboration with @mlcommons.org @eleutherai.bsky.social @jhu.edu and 97 community members, we created CommonLID, a new benchmark for LangID for 100+ languages!
10.02.2026 20:44 β π 10 π 5 π¬ 1 π 0
rdfs:subPropertyOf
rdfs:subPropertyOf
rdfs:subPropertyOf .
13.02.2026 19:14 β π 2 π 0 π¬ 1 π 0
Meta Plans to Add Facial Recognition Technology to Its Smart Glasses
Meta's plan for launching glasses that people can use to secretly identify strangers on the street is to do it "during a dynamic political environment" when people who care about why that's bad are "focused on other concerns."
www.nytimes.com/2026/02/13/t...
13.02.2026 14:54 β π 1146 π 545 π¬ 76 π 191
It may be a persons-of-a-certain-age phenomenaβ¦
13.02.2026 18:56 β π 1 π 0 π¬ 1 π 0
Great :) If i understand right, server needs to send some things as headers to lock out some dns tricks.
Are the iframe controls coming to ios/webkit et al.?
Thereβs going to be huge demand for running untrusted code sandboxedβ¦
13.02.2026 18:54 β π 0 π 0 π¬ 1 π 0
It sounds like youβre talking about the βknowledge graphsβ direction here? Although wikidata is very human and context richβ¦
13.02.2026 10:48 β π 0 π 0 π¬ 0 π 0
I might have something thatβs a good fit for this. In fact it was already giving me Joost widget flashbacks.
How tightly can this thing sandbox? eg avoiding dns exfiltrationβ¦
13.02.2026 10:41 β π 1 π 0 π¬ 1 π 1
DAV Searching and Locating (DASL) Home Page
DASL is a great name for β¦ www.webdav.org/dasl/ especially in context of read-write APIs to data stores.
Re Solid-esque APIs. Might be good to get Lisa Dusseaultβs perspective, having worked on both
www.oreilly.com/library/view...
and dtinit.org/about
13.02.2026 10:34 β π 0 π 0 π¬ 1 π 0
Rescuing Solid is a poor goal. I continue to believe in the broad mission it articulates including pragmatic use of RDF (which can be nuanced and contextualized). The fundamental issue with Solid is not RDF but the idea that full realtime read-write of shared data files is important and achievable.
13.02.2026 10:26 β π 1 π 1 π¬ 2 π 0
CHI'26 Workshop on Developing Standards and Documentation For LLM Use as Simulated Research Participants
Workshop Motivation
We've extended submissions for our #CHI2026 workshop on Developing Standards and Documentation For LLM Use as Simulated Research Participants.
Submit a short position paper by Feb 20th and let's think through some thorny issues together!
sites.google.com/andrew.cmu.e...
12.02.2026 00:53 β π 9 π 5 π¬ 0 π 0
Song of the Cerebellum
Is thought just motion in the mind?
I was shocked to learn 80% of our neurons are in the cerebellumβthat little bit of hindbrainβvs ~20% in the big fancy neocortex. (Like learning ~55% of our cells are non-human!)
Evolved perhaps for motor control/coordination, the cerebellum became the controller for many abstract mental activities.
03.02.2026 19:28 β π 6 π 3 π¬ 3 π 0
βDigitizingβ eg via gaussian splats seems a potential middle ground, where you might mine the ever changing dreamspace for material and then assemble it into a more repeatable experience. Maybe better for backgrounds etc where physics and animation expectations might be lower?
04.02.2026 19:42 β π 0 π 0 π¬ 1 π 0
X Posts about London Crime - large growth in number of likes after monetisation introduced
Wowsers look at what happened to posts on Twitter/X about London crime after monetisation of posts was introducedβ¦
economist.com/britain/2026...
31.01.2026 18:07 β π 1223 π 549 π¬ 17 π 72
Iβd suggest @cloudflare.social have a free internet service that might be able to help you, exceptβ¦.
28.01.2026 07:18 β π 0 π 0 π¬ 0 π 0
FOAF namespace server is back up at xmlns.com. Apologies for the downtime.
23.01.2026 23:39 β π 5 π 0 π¬ 0 π 0
Our paper on the mysterious Devonian organism Prototaxites has now finally been published! See the paper here (www.science.org/doi/10.1126/...) and our explainer thread below!
Prototaxites reconstruction by Matt Humpage
21.01.2026 19:25 β π 1084 π 403 π¬ 37 π 106
It's cool that Chrome's new renderer/rasterizer (Graphite) is using Dawn (Chrome's WebGPU implementation) as its GPU abstraction layer.
"Dogfooding" it like this means they're going to be extra-serious about perf/features/etc.
blog.chromium.org/2025/07/intr...
09.07.2025 19:41 β π 1 π 1 π¬ 0 π 0
Great video & project! Like many I was drawn into the SDF universe via Shadertoy, & as a standardsy person initially weirded out by each scene/model bundling its own tiny renderer. Do you hold any hope for SVG-ish interop amongst SDF tools & communities? Do you think of your units as metres, for eg?
14.01.2026 06:33 β π 0 π 0 π¬ 0 π 0
Stephen Collins comic on Twitter addiction
@stephencollins.bsky.social made the point best
13.01.2026 08:11 β π 416 π 193 π¬ 3 π 6
Jules Verne was a prophet: en.wikipedia.org/wiki/The_Pur...
07.01.2026 12:37 β π 3 π 3 π¬ 0 π 1
Plasticine dog Gromit lies slumped against a (plasticine) kitchen cabinet, holding a milk bottle. Trapped up to his neck in the bottle is Feathers McGraw, notorious criminal penguin
Well, itβs the day after Wallace and Gromit rode again β and itβs the 31st anniversary of the first appearance of the greatest sound effect in the history of the motion picture, so hereβs the story of how (I found out how) they made it...
26.12.2024 10:59 β π 504 π 182 π¬ 11 π 33
My colleague Sanjay Ghemawat & I have done a fair bit of performance tuning of various pieces of code. We wrote an internal Performance Hints document ~2 years ago as a way of identifying some general principles & we've recently published a version of it externally.
Doc: abseil.io/fast/hints.h...
19.12.2025 22:25 β π 130 π 25 π¬ 2 π 0
Loving this. "Antirender" engine converts architects presentation drawings into the likely reality of the designs :)
antirender.com
15.12.2025 12:52 β π 279 π 108 π¬ 7 π 20
Doubling every century is also exponential
17.10.2025 09:18 β π 0 π 0 π¬ 0 π 0
Iβm half Chinese, but the most tiresome type of person on here wants me to apologize to asians for calling an anti-toxicity feature βzen mode.β
03.10.2025 04:19 β π 274 π 14 π¬ 31 π 8
The flute is older than the wheel.
A flute fashioned from the bone of a vulture was discovered in a cave in Germany and thought to be 35,000 years old, making it tens of thousands of years older than the wheel.
Donβt reinvent the woodwinds.
15.09.2025 19:39 β π 497 π 88 π¬ 14 π 7
βmaking improvements to the software's qualityβ #busted #breaking #latecapitalism
16.09.2025 04:36 β π 1 π 0 π¬ 0 π 0
29.08.2025 15:41 β π 51 π 6 π¬ 1 π 0
Musings on digital marketing and what else is next | Dad β€οΈ Windsurfing | π³π± Dutch
digital daughter, learning every day π at protocol native. made by @hailey.at. i try to be a good person
my notes & blog: greengale.app/penny.hailey.at
my website: sites.wisp.place/did:plc:jv5m6n4mh3ni2nn5xxidyfsy/home
A protocol for a humane, privacy-focused, self-authenticated social web. Recipient of NGIPointer grant, graduate of Oxford Foundry, audited by Cure53 and Radically Open Security
https://peergos.org
https://github.com/peergos/peergos
Make yourself the 'Go To' person in any industry and earn a side income from running small networking events.
Earn through entrance fees, membership fees, sponsorships & deals. GetLaunched.org
They who would not sacrifice their own souls to save the whole world, are, as it seems to me, illogical in all their inferences, collectively. - CS Peirce
Full Prof. Γstfold University College, Lab http://nichele.eu/lab.html. I study bio-inspired AI with ALife & Complex Systems #CA #evo. Proud father of 2.
Born at 310 ppm.
Love Bristol, boats & wine.
Un autre monde est possible.
Green, anti-fascist, Cornish, European.
No walls, no borders. #NoPasaran! #SlavaUkraini
Val de Loire
#OpenData #LinkedData #DigitaalErfgoed #DigitalHeritage #GoudaTijdmachine #GoudaTimeMachine #genealogie #genealogy
https://toot.community/@coret https://genealogie.coret.org/
hybrid AI | neuro-symbolic AI | knowledge architecture | knowledge graphs | ontology modeling | structured content | information architecture | practice democratization | community building
Senior Associate Professor in Computer Science at LinkΓΆping University β’ Amazon Scholar β’ research on data management topics related to graph data and data on the Web
https://olafhartig.de
Evolutionary biologist & paleontologist. Website: http://bit.ly/JTCsite Macroevolution, phylogenetic comparative methods, diversification, phenotypic evolution of π πΏπ¦ Postdoc @iDiv in Leipzig
Data Strategy & Design Leader
Passionate about the power of data to help organisations make a positive difference in the world. Director at Epimorphics Ltd, former Deputy Director for Data at Defra. Based in South Wales, UK. With #aphantasia
Postdoc investigating the plants of the past at the University of Edinburgh
I help ambitious people raise capital, get board seats, take companies public. Let's get more entrepreneurs and builders on bsky!
Author of 4 best selling business books
Building a game/engine based on signed distance fields
~ Past: rendering lead @figma.com, indie VR dev, founder @ Workflowy, search backend @ Google, Quake modder
miketuritzin.com
peeking into the post-browser web
~ https://metafluff.com
~ https://webtransitions.org
~ https://userandagents.com
~ @intenttoship.dev
past lives: Mozilla, Protocol Labs, Sub Pop Records
prototypes @wikipedia.org
London
todepond.com