AI News — Wednesday, 26 August 2026
- Sam Utteker
- 1 day ago
- 5 min read
5-minute update on today's AI news
In a Nutshell
Apple launches M6 and M5 Ultra chips with massive AI compute gains, while Nvidia's earnings loom over the market. Privacy concerns intensify as Microsoft's AI watermarks are found traceable to user IDs and Anthropic's Claude starts sharing memory across services. The AI funding landscape shifts with Hugging Face fielding $13B acquisition offers, and a Stanford study confirms AI is displacing entry-level workers at alarming rates.
Apple introduces M6 and M5 Ultra chips for a big leap in AI compute
Apple's new M6 and M5 Ultra chips bring significant AI compute improvements to the Mac lineup, positioning Apple as a serious on-device AI inference contender. The M5 Ultra powers the new Mac Studio, while the M6 debuts in the Mac mini, bringing neural engine performance that rivals dedicated AI accelerators for professional workloads. Source: Apple Newsroom
Claude and Cowork now share user memory with no option to separate services
Anthropic has enabled shared memory between Claude and its Cowork product with no opt-out, raising significant privacy concerns for enterprise users. Claude Code remains separate for now, but the move signals a shift toward cross-product data sharing that organizations must factor into their AI governance policies. Source: The Register
Microsoft AI watermarks in Paint and Photos are linked to user IDs, researcher finds
A security researcher has discovered that Microsoft's AI-generated content watermarks in Paint and Photos encode identifiable user information, undermining the privacy promise of content provenance systems. This finding could influence how regulators view AI watermarking mandates under the EU AI Act and similar frameworks. Source: The Register
SpaceX claims it will put an Nvidia Vera Rubin NVL72 rack in orbit next year
SpaceX announced plans to launch a full Nvidia Vera Rubin NVL72 rack-scale AI system into orbit by 2027, with significant scale by 2028. While ambitious, the claim underscores the emerging convergence of space infrastructure and AI compute, with implications for edge AI and satellite-based data processing. Source: The Register
EPA drops requirement for public notice of polluting AI data centers
The EPA is eliminating the requirement for public notification about polluting data centers, shifting oversight to state and local regulators. As AI infrastructure expands rapidly, this rollback removes a key transparency mechanism for communities affected by data center emissions. Source: The Register
Australia's record industry refuses to chart songs recorded in the key of AI
Australia's recording industry association announced it will not chart music created with AI, effectively excluding AI-generated tracks from official music rankings. The policy mirrors broader global tensions between human creative industries and AI-generated content, with potential ripple effects across entertainment sectors. Source: The Register
Hugging Face reportedly in talks to be acquired for $13 billion
Hugging Face, the open-source AI platform that hosts millions of models, is reportedly fielding acquisition offers valuing it at around $13B. The founders' sense of responsibility to the community raises doubts about whether a sale will materialize, but the valuation signals massive demand for AI infrastructure companies. Source: TechCrunch
AI is hitting entry-level jobs hardest, Stanford study finds
A Stanford study reveals that young employment in AI-impacted fields is down 19% compared to more AI-resistant occupations, confirming that AI displacement is concentrated at the entry level. The findings have major implications for workforce development, education curricula, and hiring strategies across industries. Source: Ars Technica
Nvidia's dependence on hyperscalers faces big test in earnings report
Nvidia reports Q2 earnings on August 26, with options markets pricing a potential $280 billion stock swing. The results will test whether AI infrastructure spending by hyperscalers continues to sustain Nvidia's record growth, making this one of the most consequential earnings reports for the entire AI sector. Source: CNBC
Who's behind the new stealth AI model Ox Alpha?
A mysterious new AI model called Ox Alpha has driven parts of the AI community into a frenzy of speculation about its origins. The model's capabilities and unknown provenance highlight the growing trend of anonymous or stealth AI releases, raising questions about accountability and safety in model deployment. Source: TechCrunch
Is it legal to train AI models on copyrighted books? It's complicated
A deep dive into the legal landscape of training AI on copyrighted books reveals that most published authors have contributed to AI development without consent. The analysis comes amid a wave of copyright lawsuits that could reshape how AI companies source training data. Source: TechCrunch
DeepMind alumni's AI agent outperforms Anthropic and OpenAI at replicating research
British AI lab Inherent released Faraday, an AI agent that can replicate scientific papers, outperforming Claude and GPT at research reproduction tasks. This represents a significant step toward AI systems that can autonomously validate and extend scientific work, with implications for research acceleration and methodology. Source: TechCrunch
Harvard's $699 startup bootcamp offers AI avatars of its instructors
Harvard Business School's Foundry program uses AI avatars of real instructors to provide feedback during practice pitches and board meetings. At $699, the program democratizes access to Ivy League pedagogy, but raises questions about the role of AI-mediated instruction in formal education. Source: TechCrunch
Anthropic publishes position on open-weights models as EU AI Act looms
Anthropic CEO Dario Amodei laid out the company's stance on open-weights models, balancing the benefits of openness with safety concerns. The position comes as regulators worldwide debate how to classify and govern open-weight AI systems under emerging legislation. Source: Anthropic
Google announces Gemini 3.7 Flash just three weeks after previous release
Google released Gemini 3.7 Flash only three weeks after the 3.6 Flash debut, claiming substantial improvements. The accelerated cadence reflects the intensifying model competition, but raises questions about whether the rapid iteration pace is sustainable or adding diminishing returns. Source: Ars Technica
Become a Fellow at university-365.com
Become Superhuman. In a world of AI, the most dangerous thing you can do is skip AI training. Prompt Smart, Prompt UP!
Comments