A court has preliminarily approved Sony's $7.85M settlement over claims it monopolised digital games by killing third-party game vouchers in 2019.

Norway will bar children aged 6 to 13 from using generative AI tools in school from late August. Teens 14-16 can use them only under teacher supervision.

Aether AI, founded by UCSD researcher Biwei Huang, has raised $20mn to build causal world models for robots, a direct challenge to AI's scaling orthodoxy.

HyperLight has raised an $80mn Series C led by MediaTek to scale its thin-film lithium niobate photonics for the optical interconnects AI data centres need.

France has mobilised another €13bn through its Tibi initiative, the insurer-backed model Europe keeps studying, now aimed at funding scale-ups continent-wide.

Two campaigns this week, TeamPCP's 1,000 poisoned open-source packages and the abuse of Claude's own chat feature, show trust is the new attack surface.

John Edwards has resigned as the UK's information commissioner after admitting 'inappropriate' humour, the first such resignation in its 40-year history.

Google's AI Overviews repeatedly presented SCP Foundation horror fan-fiction as fact, Futurism found, in at least 20 cases, with no fiction disclaimer.

Ahead of its IPO, OpenAI has hired Dean Ball, the architect of Trump's AI Action Plan, to lead a new policy team, days after poaching Google's Shazeer.

A new Haberman & Swan book claims Trump privately mocked Zuckerberg and Bezos for currying favour, showing Musk their texts as 'first-class groveling'.

Reliance Jio has filed for what could be India's biggest-ever IPO, and used the same AGM to unveil a $15bn satellite network to rival Starlink.

Miami startup Subquadratic claims its SubQ model breaks the 'quadratic attention' bottleneck. Independent tests back much of it, but doubts remain.