Google deleted the open source printer driver
Also, EA sold the company and AI got more paperwork.

The IT Department Replaces the Water Cooler

Google, the company that once asked everyone to "Do the right thing" but then deleted the original sign, is now making good on its promise to restrict how people install software on their own personal Android devices. A new developer registration decree requires that all applications installed on certified devices, even when sideloaded, must come from a developer who has verified their identity with Google. The open source application distribution project, F-Droid, immediately flagged the issue, noting the policy is fundamentally incompatible with its operating structure.

Board member Marc Prud'hommeaux explained the new requirement essentially kills F-Droid, because the project cannot compel its developers to register with Google or seize their application identifiers. Google claims the rule is a security measure to protect users from the 50x higher malware rates in sideloaded apps, a statement F-Droid dismisses as a pretense for consolidating power and tightening control over an ecosystem that was formerly open. The analogy is simple: the building manager is installing a new mandatory key card system for all office doors, but only cards printed by the building manager will work, preventing the residents from setting up their own mail drop boxes.

Anthropic Quietly Upgrades The Intern, Names It Sonnet 4.5

Anthropic, the AI company with the sensible-sounding name, has announced an update to its LLM lineup with the new Claude Sonnet 4.5 model. Reports from early users suggest this middle-tier model is performing as well as the previous flagship, Opus, for common tasks like coding and deep research, but at a more affordable price point. The company is also touting improved Structured Outputs, which means the model is now better at consistently spitting out JSON formatted data without a lot of extra conversational nonsense attached to it.

This improved reliability is being hailed by some developers as "a big unlock" for agentic workflows. Other, more tired developers have noted the feature is essentially Anthropic finally catching up to the competition, but congratulations are still in order for completing this very basic task. It is all part of the continuous performance review cycle where last year's genius becomes this year's perfectly adequate mid-level manager.

EA's 36-Year Public Run Concludes in $55 Billion Private Equity Finale

Electronic Arts, the video game publisher known for its sports titles and extensive microtransaction policies, is being taken private in a massive $55 billion leveraged buyout. The acquisition is led by a consortium that includes the Saudi Arabian Public Investment Fund (PIF), which was already a stakeholder in EA, along with Silver Lake and Affinity Partners. Affinity Partners is the firm run by Jared Kushner, former Senior Advisor to the President.

The deal is being referred to as the largest private equity buyout in history and will end the company's 36-year run on the public market. The general sentiment across the industry is a weary shrug, treating the news as a formality rather than a shock. The only real question is what level of cost-cutting will commence when the spreadsheet people take over from the development people. EA CEO Andrew Wilson, however, is expected to remain in his current position, proving that even a $55 billion transaction cannot disrupt the C-suite org chart.

The New Compliance Form: California's AI Transparency Act

California Governor Gavin Newsom has signed Senate Bill 53, the Transparency in Frontier Artificial Intelligence Act, into law. The legislation targets "large frontier developers," a term that bureaucratic shorthand for companies with over $500 million in revenue and models trained using a truly absurd amount of computing power (more than 10²⁶ operations). The new law requires these large developers to create, implement, and publish a public "governance framework" on their websites, detailing their safety standards, risk protocols, and cybersecurity measures.

The government is essentially treating the creation of world-altering, potentially catastrophic AI systems as a mandatory quality assurance process, complete with a new requirement to report critical safety incidents within 15 days. California hopes that an extra layer of paperwork will contain the next generation of superintelligence, which is a very human way of handling an extinction event.

Briefs

• Open-Source Monitoring: The iRobot Founder, Helen Greiner, has stated that we should not believe the AI and Robotics hype, which is the kind of honesty you only hear after a founder is safely outside the company. She probably just wanted the Roomba to clean the corners better.
• Bureaucratic Mishap: The FCC accidentally leaked iPhone schematics, potentially giving rival companies a sneak peek at Apple's secrets. This is what happens when you let the intern manage the server folder permissions.
• Existential Threat: Oracle Chairman Larry Ellison suggested that citizens will be on their "best behavior" amid nonstop recording and AI surveillance. It is reassuring to know that the new corporate motto is essentially, "Smile, you're on camera, or we will delete your credit score."