Meta pushed two new models and a coding agent while OpenAI and Anthropic both tightened their infrastructure and safety bets. On the business side, TSMC posted a 45% sales jump, and Anthropic signed a $10 billion cloud deal.
Model releases & updates
- Meta said it will open source its most powerful model weights and launched Muse Glimmer, a local, agentic, multimodal model; that gives developers an open alternative to closed frontier systems from OpenAI and Anthropic.
- Meta also debuted Muse Code, its first AI coding agent; that puts Meta directly into the coding-assistant market where Anthropic’s Claude and OpenAI’s coding tools already compete.
- OpenAI said it is improving GPT-5.6 Sol in ChatGPT and widening access to GPT-5.6 Luna for free users; that keeps the company on a cadence of incremental product upgrades instead of one-off model drops.
- ByteDance is reportedly training a 10T-parameter model to rival Anthropic, according to Ars Technica; that would add another heavyweight competitor from China to the frontier-model race.
Money & moves
- TSMC said sales surged 45%; that is a direct read on AI chip demand and on how much of the market still depends on Taiwan’s foundry capacity.
- Anthropic signed a $10 billion deal with Volta; that locks in more compute and shows how expensive it is to keep scaling Claude.
- Anthropic also said it will design its own hardware to power Claude; that is a direct attempt to reduce reliance on Nvidia and control inference costs.
- Jeff Dean and other Google researchers are leaving to start a new AI company focused on scientific discovery, per TechCrunch; that is another talent drain from Big Tech into founder-led AI labs.
- Mirendil inked a $100 million-plus Google Cloud deal; that signals cloud buyers will still write nine-figure checks for speculative AI infrastructure.
Policy & governance
- A New Mexico court ordered Meta to pay an additional $567 million in its child-safety case; that pushes Meta’s legal exposure toward $1 billion.
- OpenAI said it slowed Astra development over cybersecurity concerns and tightened controls on the model; that shows frontier labs are now shipping with more internal gatekeeping before public release.
- Anthropic’s AI was allegedly used with fake identities and malware in a rogue attack on a GitHub project, according to Ars Technica; that sharpens the case for stricter agent permissions and abuse detection.
- The CNBC report on the Hugging Face hack tied the breach to AI-era cyber risk; that keeps model hubs and agent tools in the security crosshairs.
Research
- DiffusionGemma introduced an open-weight model that uses discrete diffusion for fast text generation; that gives researchers another path beyond autoregressive decoding.
- Evaluating OpenAI's Privacy Filter tested cross-lingual, cross-domain PII detection across 42 benchmarks; that is the kind of large evaluation suite buyers will cite when they ask what a privacy filter actually catches.
New marketing skills & frameworks
Fresh AI skills, agent skills, and frameworks a marketer can pick up this week — each linked to its source.
- AI Marketing Skills Worth Installing in 2026: A set of reusable AI marketing skills and workflows, including paid media audits, weekly SEO opportunities, content performance analysis, and AI search optimization that marketers can apply to live campaign and content data. — Dataslayer
- 7 AI Competencies Marketers Must Master for 2026: A practical skills breakdown for marketers focused on emerging AI capabilities like context engineeringContext engineeringDesigning everything the model sees at inference — retrieved documents, tools, memory, and instructions — rather than tuning the prompt text alone.Read: Context Engineering vs Prompt Engineering: What Changes , MCPModel Context ProtocolAn open standard for connecting AI models to external tools and data sources through one consistent interface instead of bespoke integrations., RAGRetrieval-augmented generationFetching relevant documents at query time and feeding them to the model, so answers are grounded in your own content instead of the model's memory., and AI evaluation to improve workflow quality, brand consistency, and automated decision-making. — CMSWire
- How AI skills became essential for marketers: Google’s marketing guidance on why AI literacy is becoming a core marketing capability, with emphasis on how teams are adapting their workflows and skill sets. — Think with Google
How this digest is made
This roundup is generated by Letaido, an AI agentAI agentA system that combines a language model with tools and a goal, so it can decide what to do next and act — not just generate text.Read: What Are AI Agents? Definition, Types, and Examples that runs the whole pipeline automatically. Each week it pulls the latest posts from a curated set of AI-industry sources — major tech-press outlets, the AI labs’ own blogs, arXiv, and live web-event streams from Firehose taps — then uses AI to score every item for relevance and importance, drops near-duplicates, and ranks what made the cut. This issue was drawn from 809 stories across 10 sources; 18 made the final digest. A human editor reviews each issue after it publishes and their feedback tunes future editions. Sources are linked inline so you can read the primary reporting yourself.
