In their letter defending open-weight AI, tech companies urge lawmakers to avoid "premature restrictions" that "stifle competition or drive innovation overseas" (Ashley Capoot/CNBC)
https://www.cnbc.com/2026/07/24/nvidia-microsoft-meta-ope…
🇺🇦 #NowPlaying on KEXP's #SundaySoul
Luther Vandross:
🎵 Never Too Much
#LutherVandross
https://ericfaria.bandcamp.com/track/luther-vandross-never-too-much-eric-faria-remix
https://open.spotify.com/track/3nFJbZCHP4d9vduKjJLdBL
Institutional Data. Open Knowledge. AI. How can we create better bridges between them and work together to solve some of our big issues around data and access to knowledge in the age of AI?
The coming AI-BRIDGES Symposium in London (May 28-29) looks at these questions.
If you care about Open Knowledge, Institutional Data and AI, join us!
Registration open till May 21st. It's free, but seats are limited. Link to full program and registration here:
Just over a week after Mosaic 1.0 was released, CERN gifted the Web to the world as open source software. As of 30 April 1993, the still relatively new Internet communications platform was suddenly free for anyone to use, with no strings attached.
https://cybercultural.com/p/1993-mosai
It can be exhausting to keep up with the breakneck pace of cybersecurity, so don't miss today's Metacurity for the most crucial developments you should know, including
--Coupang hit with record $409 million fine over massive insider-driven data breach,
--CISA orders faster patching as AI speeds exploitation,
--Suspected Russian hacker extradited to the US over Void Blizzard,
--OpenAI disrupts China-linked campaigns targeting US tech debates,
--Digital bre…
🇺🇦 #NowPlaying on KEXP's #VarietyMix
Luther Vandross:
🎵 Never Too Much
#LutherVandross
https://ericfaria.bandcamp.com/track/luther-vandross-never-too-much-eric-faria-remix
https://open.spotify.com/track/3nFJbZCHP4d9vduKjJLdBL
Unexpected, but not unwelcome:
Millionaires including former footballer Gary Lineker and music producer Brian Eno have written to new prime minister Andy Burnham asking to be taxed more.
In an open letter, 120 well-off Britons told Burnham: "We can afford it. We're not talking about higher taxes on those who get up and go to work for their income every day, but on the very richest whose income is derived from the wealth they hold."
Organised by Patriotic Millionaires, the letter said it would lead to a more equal society, and urged a "devolution of wealth and power from the very richest".
#ukpol
Decode-Time Grammars: Constrained LLM Generation over a Refinement Order of Grammar Fragments
Shuoming Zhang, Ruiyuan Xu, Haofeng Li, Qiuchu Yu, Yangyu Zhang, Chunwei Xia, Xiaobing Feng, Chenxi Wang, Huimin Cui, Jiacheng Zhao
https://arxiv.org/abs/2607.18357 https://arxiv.org/pdf/2607.18357 https://arxiv.org/html/2607.18357
arXiv:2607.18357v1 Announce Type: new
Abstract: Large language models now write a growing share of the world's code, increasingly inside agents and serving systems that compile, execute, or dispatch generated code without line-by-line review. This works well for mainstream languages but remains brittle for low-resource programming surfaces such as domain-specific languages, custom library APIs, and command-line tools. Even under grammar-constrained decoding, a model can still produce references invalid in the current environment: a buffer never declared, a column absent from the schema, a function the library does not provide, or an unsupported CLI option.
This paper introduces decode-time grammars: grammar fragments instantiated during generation from a runtime environment Gamma. A region-specific policy selects a fragment for each hole, and a tightening operator replaces open reference positions with Gamma-typed slots whose candidates are exactly the names, fields, APIs, or options available at that point. Newly generated declarations enter Gamma before later regions are decoded, so the constraining grammar can depend on the prefix already generated. This ensures not only grammatical correctness but also semantic correctness, by preventing references to undefined symbols.
We formalize grammar fragments as environment-indexed grammars ordered by refinement, prove No-Ghost soundness for Gamma-slotted fragments, show that refinement preserves this support-set guarantee, and characterize the boundary of mask-enforceable properties. We implement the approach in gproj with offline grammar induction and online policy resolution. Across TileLang, SQL, and P4, with models from 0.6B to 236B parameters, gproj eliminates ghost references by construction at moderate overhead over standard constrained decoding.
toXiv_bot_toot
OpenAI Chief Futurist Joshua Achiam is leaving the company later this month after nearly nine years (Maxwell Zeff/Wired)
https://www.wired.com/story/openai-chief-futurist-joshua-achiam-is-leaving-the-company/
🧵 1/N
I've always thought it was a bit ironic that git was designed to be perfectly decentralized, only to have Github become the one centralized way one uses git (for many people, at any rate).
But this post makes a good point: the centralization also means that it functions as a reliable archive. If a popular open source project spins up its own git server, and then later goes mostly defunct -- or least, the git server goes away -- the code is pretty much gone. But with git…