| ZOTPAPER |
| |
| News without the noise |
| Friday 9 October 2026 · The Daily Edition |
| |
From the editor A week of reckoning for AI agents: from OpenAI's creative output to AWS and Microsoft's new guardrails, while Australia's biggest datacentre IPO stumbles and the government writes the rulebook. |
Business · 3 sources · 2 min
AI datacentre company Firmus Technologies is sharply cutting its near-$44bn valuation or may shelve its initial public offering entirely, multiple sources briefed on the matter have told Guardian Australia, just weeks out from its anticipated ASX listing.
Why it leads Firmus's dramatic pre-IPO valuation cut signals a cooling in Australia's AI datacentre boom, a story with direct market momentum and implications for infrastructure investors. |
| |
AI · Developing · 4 sources · 2 min
OpenAI has released 377 newly solved mathematical problems covering algebra, number theory, theoretical computer science, mathematical logic and topology, stunning mathematicians with the accelerating capabilities of artificial intelligence. The findings, released earlier this week, follow last month's announcement that the company cracked the Navier-Stokes equation, one of the Clay Mathematics Institute's seven Millennium Prize Problems and accompanied by a $US1 million reward offer.
Why it leads The sheer volume and speed of OpenAI's mathematical output reopens the debate on machine creativity, and with 377 newly solved problems, the story is accelerating. | | | AI · 2 min
Anthropic has released Claude Haiku 5.5, the first update to its smallest and most affordable model in nearly a year, slashing per-token prices by up to 90% for smaller requests while claiming dramatic improvements across agentic coding, computer use and reasoning benchmarks compared with its predecessor.
Why it leads Anthropic slashes inference costs by up to 90%, a price war that reshapes the economics of deploying small language models and demands attention from anyone building with AI. |
| | Editor's picks AI · AWS's open-source sandbox gives developers a practical tool to constrain autonomous agents, addressing a critical safety gap as agent use grows. AI · Australia moves to codify AI safety with rules akin to banking and aviation, a legislative framework that could set a precedent for other markets. AI · Microsoft's Copilot gains direct OS control and file access, a leap in agent capability that blurs the line between assistant and autonomous operator. Security · OpenSSH deliberately disables compression and blocks username characters to fix vulnerabilities, a rare trade-off that highlights the severity of recent SSH bugs. |
| | Signal (beta) · what the people building AI said
Copilot reaches across Windows
Microsoft says Copilot will, with permission, use PC context, take actions across Windows and call local models under its Hybrid Intelligence approach. Replit CEO Amjad Masad says desktop AI apps can expose users to supply-chain attacks and severe agent errors, and says Replit is focusing on security and reliability.
With your permission, Copilot can tap into the context on your PC, take action for you, and use local models when it makes sense
Satya Nadella, Microsoft CEO · On X
Desktop AI apps are great but expose users to supply-chain attacks and catastrophic mistakes by agents.
Amjad Masad, Replit CEO · On X
|
Today's segment · Source Spotlight
The Guardian
Credibility 9/10 on the desk's rubric. 293 of this week's stories drew on it. |
|
|
| |
Unsubscribe · Manage preferences · View in browser
zot.news · Privacy · editor@zot.news
You're receiving this because you subscribed to the Zotpaper Daily at zot.news.
Zotpaper · Brisbane, Australia
|
|