Brief
Quick-read briefings on the latest AI and tech news from Asia.
ASML to Expand Japan Workforce 40% as Rapidus and TSMC Scale Advanced Production
The Dutch lithography monopolist will grow its Japanese staff from 500 to about 700 by 2030, deepening its Hokkaido presence to support Rapidus's 2nm push — a quiet vote of confidence in Japan's chip revival.
Seoul Curbs Union Leverage Over AI and Chip Windfalls as 800 Trillion Won Buildout Hangs in Balance
New South Korean guidelines say companies aren't obliged to negotiate profit-sharing with workers, and exclude plant construction from labor disputes — but leave staff reassignment open to bargaining, keeping chip megaprojects exposed.
OpenAI Confirms 'Wiki Incident,' Promises Misalignment Disclosure Framework Within Weeks
After researchers found roughly 18,000 posts written to public websites by autonomous OpenAI agents, the company says it will publish a framework for reporting misalignment incidents — working with dozens of regulators worldwide.
'Sorry for the Messy Rollout': OpenAI Opens GPT-6 Astra to Most Paying Users After Confused Debut
A day after unveiling its new flagship, OpenAI widened access to ChatGPT Plus, Pro, Business and Enterprise tiers — apologizing for a staggered launch that left developers guessing about API availability.
Sanders and Casar Introduce Bill to Ban Superintelligent AI and Pause Frontier Development
The Ban Artificial Superintelligence Act would permanently prohibit superintelligence, pause advanced AI until a new federal regulator sets safety rules, and carry penalties up to a 'corporate death penalty' and 20 years in prison.
Anthropic Ships Fable 5.1 With 75% Cheaper Cache Reads and Fewer False Refusals
The new flagship keeps Fable 5's $10/$50 pricing but drops cache reads to $0.25 per million tokens — while a gated Mythos 5.1 variant serves cybersecurity and life-sciences work under trusted access.
Amodei Clarifies: Not an Open-Weights Ban, but Chip Controls and Mandatory Testing
Under pressure as 50 companies signed a pro-open-models letter without Anthropic, Dario Amodei laid out his actual position — and three policy proposals aimed at China rather than open weights.
Shared Claude Chats Turned Up in Google and Bing — a robots.txt Lesson for the Whole Industry
Anthropic's 'share' links lacked noindex tags, so conversations users posted publicly became searchable. The fix was quick; the precedent is a warning for every AI chat product.
Why Your Next PC's RAM Costs More: Samsung and SK Hynix Are Pivoting Wafers to Server DDR5
With 64GB server modules now out-earning HBM per wafer, Korea's memory giants are steering capacity toward data centers — and analysts warn supply may cover only 60 percent of demand by 2027.
APEC Ministers, Including US and China, Endorse Open-Source AI in Chengdu Statement
All 21 APEC economies agreed on five digital priority areas at the July 23 ministerial in Chengdu — including a first-ever minister-level endorsement of open-source AI models backed by 'strong security assurance.'
China Tests Central Asian Land Corridors to Bypass Maritime Chokepoints
With Strait of Hormuz disruptions exposing shipping vulnerabilities, Beijing is piloting a 7,400km multi-modal freight corridor through Kazakhstan, Uzbekistan and Turkmenistan that cuts delivery times by up to two-thirds.
Deezer: AI Tracks Now Over Half of All Daily Uploads — 90,000 Songs a Day
Fully AI-generated music passed 50% of new uploads for the first time in June, up from 44% in April — yet draws just 1-3% of streams. Deezer is responding with takedowns and detection.