-
Linus Tovalds: Linking Patchwork with Sashiko? (Jul. 14, 2026)
“Linux is not one of those anti-AI projects, and if somebody has issues
with that, they can do the open-source thing and fork it.Or just walk away.
AI is a tool, just like other tools we use. And it’s clearly a useful one.
It may not have been that “clearly” even just a year ago, but it’s no
longer in question today.” -
Simon Willison: Claude make Fable 5 permanent (Jul. 18, 2026)
Claude will include Fable 5 in Max and Team Premium plans from July 20, at 50% limits. Pro and Team Standard users keep access via usage credits and a one-time $100 credit; $20/month plans still lack Fable 5. -
NY Times: India Is Moving Fast to Build A.I. Data Centers. A Coastal City May Pay the Price. (Jul. 15, 2026)
Visakhapatnam is becoming a coastal hub for massive A.I. data centers as Google, Adani, and Brookfield pour in investment and receive large subsidies. -
Simon Willison: xAI's Grok CLI uploaded directories (Jul. 15, 2026)
xAI’s Grok CLI uploaded directories, exposing SSH keys and password stores; uploads were disabled, retained data deleted, and grok-build open-sourced under Apache 2.0. -
Charles Azam: Fable 5 vs. GPT-5.6 Sol on an NP-Hard Problem: Does /goal Help?
Claude Fable 5 outperformed GPT‑5.6 Sol on a hard fiber‑network optimization benchmark, delivering better speed, accuracy, and consistency. -
WSJ: AI’s Wider Availability Is Good for China, Not Great for OpenAI and Anthropic (Jul. 17, 2026)
AI is turning into a cheap, widely available commodity as lightweight, and open-weight models from China, Meta, and others match top U.S. systems. That fuels price wars, squeezes OpenAI, Anthropic margins, and blurs who owns what AI learns. -
WSJ: Flash Heard: What IBM’s Woes Mean for the Rest of the AI World (Jul. 15, 2026)
IBM warned earnings would suffer as big customers shifted spending from mainframes and software to servers, storage, memory for AI, squeezing other budgets. -
WSJ: New York Set to Temporarily Ban Large New Data Centers (Jul. 14, 2026)
New York will pause construction of large data centers for up to a year, halting projects of 50 megawatts or more. Officials will write rules to protect the power grid, limit environmental harm, and boost community benefits. -
State of Open Source AI: The State of Open Source AI — V1.0 · July 2026 (Jul. 1, 2026)
Open-weight models now match closed models on many tasks, drove token routing to open offerings, and inference costs plunged dramatically, but closed models still lead on reasoning. -
NY Times: Someone Used A.I. to Write an Unauthorized Biography of Me. I Don’t Recommend Reading It. (Jul. 16, 2026)
A.I.-generated biographies are flooding Amazon, often flattering, fabricated, low-quality, and produced by prolific publishers using stock photos and fake bios. They sell modestly, exploit platform rules, and leave subjects surprised, exposed, and unable to contact creators. -
WSJ: AI Videos Are Flooding TikTok Shop (Jul. 15, 2026)
TikTok Shop’s AI tools auto-generate shoppable videos, boosting scale and affiliate income, but stirring complaints from creators and brands about misleading content, lost commissions, and erosion of consumer trust. Some brands now ban AI promos for affiliates. -
NY Times: Xi Jinping of China Pitches ‘Openness’ in Push to Shape the Path of A.I. (Jul. 17, 2026)
Xi Jinping urged open, global collaboration on artificial intelligence, promoting open-source sharing, and offering China’s help to developing countries. He cast China as a rival to U.S. firms, pushing models like Moonshot’s Kimi K3, and the World AI Cooperation Organization. -
NY Times: China’s Chip Champion to Raise Billions in Race for A.I. Control (Jul. 15, 2026)
CXMT aims to raise about $9.8 billion in a Shanghai IPO, riding surging A.I. demand for memory chips. Government backing helped growth, market share, and returning talent. -
WSJ: The AI Backlash Has Tech Executives Fearing for Their Lives (Jul. 15, 2026)
An attempted firebombing at Sam Altman’s home spurred a surge in threats, harassing visits at Anthropic, and police alerts. Firms have boosted security, hired guards, and softened warnings about AI amid public anger over jobs and costs. -
Elie Bak: J-space, open models (Jul. 14, 2026)
“Anthropic’s Verbalizable-Workspace paper showed, on one closed model family, that a model’s middle layers carry a dictionary of directions that causally steer its output.”
Leave a Reply