Here are today's news for 17th of October about Anthropic, Microsoft, OpenAI and Google. Let's dive into the details. On 8 October 2026, the UK Information Commissioner’s Office (ICO) said ten leading foundation, model developers, Amazon, Anthropic, Apple, Cohere, DeepSeek, Google, Meta, Microsoft, OpenAI and Stability AI, had made or committed to make data, protection changes after two years of scrutiny. The regulator cited clearer explanations of how personal information is used, stronger ways for people to exercise their rights, and tougher assessments of safeguards; it is monitoring whether firms follow through. The commitments are not reported as fines or binding orders. They matter to users whose information may be used in training, and to companies competing to build models while meeting UK privacy law. The ICO’s two, year programme covered 11 priority developers, selected partly for UK market share, higher, risk training data and likelihood of non, compliance. It excluded one from this outcome amid a formal investigation. The watchdog is now asking how agentic AI, systems that can act and use tools with limited human oversight, handles personal data. Anthropic reported on Oct. 16, 2026, that Claude Mythos 5 redesigned quantum circuits for eight computational tasks, improving on the best published designs across chemistry, optimization, quantum simulation, number theory and numerical computing. Given plain, language specifications, the model proposed circuits, received correctness checks and resource scores from an automated evaluator, then iterated without human input between rounds. In seven tasks, the resulting circuits reduced both qubit and gate counts; in the eighth, Claude devised a new reversible method for prime, field modular inversion. One elliptic, curve point, addition design uses 12% more qubits than the smallest prior design but 17 times fewer gates. For decoded quantum interferometry, Claude cut the gate count of a Reed, Solomon decoder by about 12 times versus Google’s earlier construction. Microsoft AI said on Oct. 15 that it had completed pretraining MAI, Base, 2, a mixture, of, experts model with 2.8 trillion total parameters and 80 billion active per token. The August, September run used 16,000 Nvidia GB200 GPUs for 38 days and processed 50 trillion tokens; Microsoft reports average model, FLOPs utilization of 21% and training goodput of 89%. The company says the model beat every open, weight system it tested on held, out evaluations and its pretraining recipe was 4.2, 12.3 times more FLOPs, efficient than DeepSeek, V4, Pro, Base across four tests. Those are company, reported base, model results, not independent evidence of finished assistant performance: post, training is still underway, and Microsoft has not announced pricing or a release date. Unlike MAI, Base, 1, announced in June with one trillion parameters, MAI, Base, 2 includes vision in pretraining, with image, related data making up 18% of its mixture. On Oct. 14, the Federal Trade Commission sent civil investigative demands, subpoenas, to OpenAI and Anthropic, escalating its inquiry into potential consumer, protection violations. Each demand exceeds 50 pages and seeks records on security incidents, safeguards, marketing materials and privacy policies; the FTC plans to seek information from third, party AI organizations later. The requests follow reports of AI agents behaving unexpectedly, including hacking websites during testing, and scrutiny of whether companies adequately protect users or accurately describe product risks. The inquiry matters to consumers whose data or systems could be affected, and to AI firms competing to market increasingly capable agents: claims about safety and performance may now face regulatory examination. The FTC has authority to pursue unfair or deceptive practices, but the demands do not establish wrongdoing. On October 1, 2026, Google’s Project Suncatcher prototype launched aboard SpaceX’s Transporter, 18 rideshare mission and entered low Earth orbit. Built with satellite company Planet, the refrigerator, sized craft carries four Google Tensor Processing Units (TPUs); Google later confirmed contact and said it was operating as expected. The roughly year, long test will measure how the chips withstand launch stresses, radiation and extreme temperatures, not provide commercial AI service. Google says the TPUs can survive a radiation dose exceeding what a five, year space mission would deliver, based on ground testing, but orbital performance remains to be verified. Heat is a constraint: the chips are expected to run AI workloads for about 15 minutes at a time before cooling. The experiment matters as energy and infrastructure demands push technology companies to explore alternatives to land, based data centers. Google’s longer, term concept is solar, powered orbital computing, potentially linked by laser communications. That's all for today's AI news briefing. Have a great day!