TechDefused Newsroom AI

AI-assisted newsroom output reviewed by a named human editor before publication.

Articles

ECB backs plan to centralise crypto supervision under EU markets regulator

The central bank warns that regulatory fragmentation under MiCA is enabling firms to shop for favourable licensing jurisdictions

Microsoft confirms 365 Copilot agent with OpenClaw-like local capabilities ahead of Build showcase

The software giant is targeting enterprise customers with a managed, secure alternative to the open source local agent tool

Vercel eyes IPO as AI agents drive deployments and annual revenue hits $340m run rate

The web hosting company says 30% of apps on its platform now originate from AI agents rather than human developers

Stanford report finds growing gulf between AI experts and public as anxiety mounts

Only 10% of Americans say they are more excited than concerned about AI, against 56% of experts who expect a positive impact

Staff cuts and bottlenecks slow US approvals for Nvidia and AMD chip exports to China

The Bureau of Industry and Security has lost nearly a fifth of its workforce since 2024, pushing licence turnaround times to double their 2023 average

Meta is building a photorealistic AI version of Mark Zuckerberg to talk to its staff

The project, reported by the Financial Times, is distinct from a separate "CEO agent" tool designed to help Zuckerberg manage his own workload

Intel Nova Lake desktop processors leaked with up to 52 cores and new LGA1954 socket

Internal documents seen by VideoCardz point to a significant performance and cache leap for Intel's next desktop platform, due late 2026

Microsoft axes Dev and Canary channels in Windows Insider programme overhaul

The software giant is simplifying its beta testing tiers, retaining Release Preview while overhauling the Beta channel

Apple releases iOS 26.5 beta 2 for iPhone

The build adds a suggested places feature to Apple Maps and continues end-to-end encryption testing for RCS messaging

Anthropic pushes back on claims of Claude degradation as user complaints mount

Viral benchmarks and a detailed GitHub analysis have fuelled debate over whether the AI developer has quietly reduced model performance

Anthropic cuts Claude Code cache window from one hour to five minutes, users report faster quota drain

The AI developer says the change should not increase costs, but forum complaints suggest some users are seeing heavier consumption

OpenAI CEO Sam Altman targeted in second home attack

OpenAI CEO Sam Altman was reportedly targeted at his San Francisco home in a second attack, with arrests and charges following, police say.

Pixel Societies: the hackathon project building pixel-art AI twins to test human compatibility

A UCL hackathon team has built a prototype that creates digital replicas of real people to simulate social interactions

OpenAI signs 88,500 sq ft King's Cross lease as first permanent London office

The US AI group will house more than 500 staff in the new space, despite pausing its UK Stargate data centre plans

Kraken launches Ink Points loyalty season

Kraken is rolling out Ink Points on Kraken Pro to reward active traders through a season-based program.

Google expands AI tools for education and research

Details of new training, NotebookLM and Gemini integrations, research partnerships and data-transfer updates in a company blog post

SoundHound AI expands agentic platform into telecom

Bitcoin safe from quantum without a fork, researchers say

Researchers say a method exists to protect Bitcoin from quantum threats without requiring a network hard fork.

Apple tests four frame styles for upcoming smart glasses

Apple is testing at least four frame designs and premium materials for a new pair of iPhone‑paired smart glasses, Bloomberg reports.

World Cup prompts wide cyber security preparations

US and allied partners are mobilising federal, state and local cyber defences ahead of World Cup matches to counter a broad range of digital threats.