Anthropic introduces Claude Fable 5.2 and Mythos 5.2, Meta releases Muse Ember, and more - Daily AI Brief #579
In this edition, covering 9 November:
- Anthropic introduced Claude Fable 5.2 and Claude Mythos 5.2, and set a price range of $70 to $78 a share for its IPO.
- Australia will make safety and security standards mandatory for frontier AI companies after an OpenAI agent accessed Medicare statistics.
- Meta released Muse Ember, its lightest and lowest cost model, for repeated agent tasks.
- Google introduced two new TPU systems, TPU 8tx for confidential inference and TPU 8i Max for million token context windows.
From our lists
A new AI Brief list covers AI product photo tools for online stores.

AI product photo tools for online stores
Anthropic and Akamai
Anthropic introduced Claude Fable 5.2 and Claude Mythos 5.2, with better coding, reasoning and vision and outputs up to 30% faster than the 5.1 models. Fable 5.2 costs $8 per million input tokens and $40 per million output tokens. Mythos 5.2 stays limited to approved cybersecurity and biomedical research programs. Fable and Mythos 5.2
Anthropic set a price range of $70 to $78 a share for its Nasdaq IPO of 825 million Class A shares in an amended prospectus filed on 10 November. Anthropic sells 801 million of the shares and existing holders sell the rest. The filing also describes a planned $26 billion private placement with AMD, Broadcom, Google, Micron, Microsoft, Nvidia and TSMC Partners. IPO price range
Morningstar estimates Anthropic's equity value at $1.7 trillion in its base case, with a bull case of slightly more than $3 trillion and a bear case of $500 billion. S&P Global Ratings cut Akamai's issuer credit rating to BB+ from BBB- because Akamai will fund the capital spending for its $11.6 billion Anthropic contract entirely with debt. Morningstar valuation · Akamai downgrade
Chips and infrastructure
Google introduced TPU 8tx for confidential inference and small scale training and TPU 8i Max for million token context windows, along with software updates for using TPUs with PyTorch. Google TPUs
Governments
The Australian Government will introduce mandatory safety and security standards for frontier AI companies after an OpenAI agent accessed non-public Medicare statistics. A new AI safety regulator will enforce incident reporting, independent model evaluations and security controls, and the government accepted all 18 recommendations of a review of the incident. Australia's AI safety standards
The UK prime minister joined a call chaired by French President Emmanuel Macron on international AI cooperation, covering AI safety, transparency, risk management and incident reporting. The UK plans to make AI a priority during its 2027 G20 presidency. Israel's National AI Directorate plans domestic AI infrastructure with about 300,000 GPUs, two gigawatts of electricity capacity and an Israeli foundation model to reduce reliance on foreign technology companies. Paris call on AI safety · Israel's 300,000 GPUs
Meta, Samsung and Adobe
Meta Superintelligence Lab released Muse Ember, which Meta calls its lightest and most cost efficient model, built for repeated agent tasks such as tool calls, coding, routing, extraction and long document parsing. It handles text, images and audio and costs 10 cents per million input tokens. Muse Ember
Meta is forming a sales organization for its AI products and building ties with consulting firms and systems integrators to reach larger companies, The Information reports; Meta has not confirmed the plan. Samsung is testing an early access Galaxy AI Premium subscription, and app code lists prices of $4.99 a month or $47.99 a year, which Samsung has not confirmed. Meta AI sales team · Galaxy AI Premium
Adobe and The Walt Disney Company formed a multi-year alliance to build custom generative AI models and creative tools for Disney's parks, cruise, studio and sports development work. Adobe and Disney
Autonomous driving
Wayve will train and evaluate its AI Driver foundation models on CoreWeave Cloud with Nvidia Blackwell GPUs under a multiyear agreement, as it prepares for commercial deployment. Tesla said on X that its robotaxi service would start in San Antonio on 9 November, without details on hours, vehicles or safety monitors; Waymo already runs driverless rides there. CoreWeave and Wayve · Tesla in San Antonio
Funding
Nvidia invested in Korean AI interconnect startup Point2 and plans investments in Motif Technologies, RLWRLD and Upstage, it said at a startup forum in Seoul. Spines raised $64 million, a $29 million Series A-1 led by Zeev Ventures plus a $35 million non-dilutive facility from Twelve, and says its revenue run rate reached $50 million. Coframe raised a $25 million Series A co-led by Khosla Ventures and Hanabi Capital for AI that creates, tests and improves websites, landing pages and ads. Nvidia in Korea · Spines · Coframe
Astral raised $5.5 million in a round led by Andre Charoo of MapleVC for its personalized AI tutor, which it says thousands of families use in more than 33 states. Snapp, co-founded by Caviar founder Shawn Tsao, came out of stealth with $4.6 million from a16z Speedrun and others for real time sports AI. Newcastle startup Crowd Threat raised £1 million for threat reports from paid local contributors in 17 countries, which AI and human reviewers verify. Astral · Snapp · Crowd Threat
Legal AI
Addleshaw Goddard will roll out Harvey to more than 3,400 lawyers and staff across Europe, the Middle East and Asia from January 2027 and develop legal AI tools with it. The American Arbitration Association expanded its AI Arbitrator to two party, document only commercial cases, where the system drafts recommendations for trained human arbitrators to review and decide. Legaltech Hub founder Nikki Shaver leaves on 1 December to become chief AI and knowledge officer at Davis Polk & Wardwell, and COO Jeroen Plink becomes CEO. Addleshaw Goddard and Harvey · AI Arbitrator · Legaltech Hub CEO
Health
NYU Langone's Lang1 models, trained on records of more than 10 million patients, predicted insurance denials, 30 day readmissions and deaths in hospital more accurately than large general models in several tests. The American College of Physicians and OpenEvidence will integrate ACP clinical guidance and Annals of Internal Medicine content into OpenEvidence's AI clinical decision support platform for US physicians. Lang1 · ACP and OpenEvidence
Also
Medium launched TK, a writing app whose AI assistant answers quick questions and pulls in notes a writer saved earlier to help turn ideas into drafts. Automattic, the owner of WordPress.com, launched Spacefast, a service that puts apps and websites built with AI coding tools online, in a market that includes Lovable. Red Hat made OpenShift 5 generally available at KubeCon North America, with agentic troubleshooting through its Lightspeed tools, a built-in MCP server and post-quantum cryptography controls. Medium TK · Spacefast · OpenShift 5
Skyfall AI launched Steve, which it calls an AI CEO, and says Steve runs three digital businesses, AlphaSocial, GammaPromo and DeltaLeads, without humans in the operating loop. Banking AI company Glia says it passed $100 million in annual recurring revenue in the third quarter while staying cash flow positive, with AI product sales up 93% year on year. Skyfall Steve · Glia
We hope you enjoyed this article
Consider subscribing to one of our newsletters like Daily AI Brief.
Also, consider following us on social media:
Daily AI Brief
Daily report covering major AI developments and industry news, with both top stories and complete market updates