Issue #99
Anthropic's IPO Filing, Sonnet 5.5, and Meta's Enterprise Push
Tuesday, September 29, 2026
Today's digest covers Anthropic's prospectus disclosures and its release of Sonnet 5.5, safety-related developments at OpenAI and in frontier AI training, and product and platform moves from Google and Meta. Also included is an Australia-focused update.
1 of 7 · OpenAI
How we will do better for Australia
OpenAI issued an apology for incidents involving Australian government websites and detailed new safeguards aimed at strengthening the country's cyber defences. The company also outlined additional support measures for Australia's cybersecurity efforts.
Read the full story →2 of 7 · OpenAI
Towards safety cases for frontier AI training
A new set of early guidelines proposes how developers can build safety cases for frontier AI training, covering technical safeguards, operational practices, and methods for investigating misalignment incidents. The guidelines aim to establish structured approaches to demonstrating safety during the development of advanced AI systems.
Read the full story →3 of 7 · TechCrunch AI
Anthropic’s prospectus details losses, growth, and, yes, a warning that its AI could end humanity
Anthropic's investor prospectus disclosed that the company is losing tens of billions of dollars annually while experiencing rapid growth. The document also included a warning that its AI could pose an existential risk to humanity.
Read the full story →4 of 7 · TechCrunch AI
Anthropic releases Sonnet 5.5, which it calls a significantly cheaper, faster work partner
Anthropic has released Sonnet 5.5, the latest version of its mid-range model, featuring faster response times and reduced token usage. The company positions the update as a cheaper, faster option for work tasks.
Read the full story →5 of 7 · TechCrunch AI
OpenAI reportedly ditches model over safety concerns
OpenAI reportedly scrapped a model after it showed a poor ability to follow instructions, according to a top executive who spoke to the Wall Street Journal. The executive cited safety concerns tied to the model's behavior.
Read the full story →6 of 7 · TechCrunch AI
Google is killing off Gemini’s Gems in favor of ‘skills’
Google is discontinuing Gems, a Gemini feature that let users build task-specific agents, and replacing it with a new capability called skills. The change comes as competitors including Meta push broader, all-in-one AI agents.
Read the full story →7 of 7 · TechCrunch AI
Meta launches enterprise AI platform, hires MongoDB CEO to lead new initiative
Meta announced an enterprise AI platform that will bring its technology stack, including Muse, Meta Business Agent, Muse API, and Muse Code, to businesses and developers. The company also hired the CEO of MongoDB to lead the new initiative.
Read the full story →That is today's roundup. Follow the links for full coverage, and see you tomorrow morning.
Morning Roost Issue #99