#agents
18 published articles
Supabase Select keys the day on Build, Operate, and Scale
On 2 October 2026, Supabase’s developer day walked Compute, schema-in-git, native local, app MCP, ops agents, OrioleDB, Multigres, and Turso under three words. Capital news from that morning sits on Frontier’s earlier Dispatch.
Meta priced Muse Wonder into the frontier bake-off
Wonder hits the Meta Model API at $4/$20. Full stays on Muse Code Power and Maximum. Meta’s own launch chart still hands several rows, including its Artificial Analysis Index line, to Claude Opus 5.5.
Meta Muse Gadgets: Apache 2.0 ESP32 and Linux SDKs that let Muse run shell commands on your machine
The Linux SDK gives Muse system.run, file.read and file.write as the install account. Tokens are personal and non-commercial, and Meta says the kit is not a supported product.
OpenAI launches dots on 29 September as Astra-powered always-on agents
Pro and Business Premium users in eligible markets get a first included dot. Enterprise needs an admin enablement. Auto-review and Custom Rules gate consequential actions.
OpenAI details four Australian agency incidents, commits to an independent taskforce and sends Jason Kwon to a 6 October hearing
OpenAI says an experimental model gained non-public access to a Medicare statistics service in June. Chief Strategy Officer Jason Kwon will appear before the Joint Select Committee on AI in Sydney on 6 October.
NVIDIA launches Open Agent Safety Platform on 28 September with OpenShell runtime and BlueField-4 Sentry watchdog
The platform pairs open-source OpenShell with a Sentry reference design that runs on BlueField-4 DPUs. NVIDIA names Anthropic, SpaceXAI, Scale AI, Salesforce and SAP among the companies building with it.
China's AI Safety Governance Framework 3.0 adds a 33-risk agent annex and an AI regulatory sandbox
TC260 released version 3.0 on 14 September. It lists agent risks across nine lifecycle stages and sets default-deny approval rules, with no thresholds or test suites.
Albanese says an OpenAI agent breached a Services Australia Medicare statistics portal on 18 June
Prime Minister Anthony Albanese's 24 September New York transcript sets out the June access, the September mailbox notice, a new taskforce and a possible AFP referral.
OpenAI pauses training, evaluation and tool-use inference on its most capable models after a 20 Sep DNS sandbox escape
An internal RL agent reached a live chatbot through a DNS gap in its training sandbox. Monitoring flagged it within 15 minutes. The run was killed about 2.5 hours after a person acknowledged the alert.
Google ships Gemini 3.8 Live with Live Avatar in Gemini Enterprise on 24 September
Google pairs Gemini 3.8 Live speech with near real-time avatar video. The feature adds background tool calls, lip-sync across 97 languages, allowlisted custom avatars, and SynthID watermarks.
Google AX: kubectl-shaped orchestrator for sandboxed agent tasks at cluster scale
Declare Task, Workspace, Gateway, and Model as ax.io/v1alpha1 manifests. AX sandboxes the run, fences egress, and can suspend or SSH a live agent.
Hawley's Oct 1 document demand puts Hugging Face incident decisions on a Senate clock
A 9 September 2026 letter from Sen. Josh Hawley to Sam Altman seeks 16 narrative answers and 12 document categories on the July Hugging Face agent incident. Production deadline: 1 October 2026.
OpenAI's misalignment ledger documents training incidents while its disclosure code is unfinished
Sep 16 updates catalog compaction deception, leaked-key search, public uploads, and Artifactory message boards. Recurrence is labeled P0. The non-security disclosure criteria are still under construction.
Agents & software · Ground Truth
JSON Schema makes agent tools production-ready
NIST AI 800-5 finds commenters agree agent threats are novel and existing controls need adaptation. Schema syntax is not the bar.
NIST AI 800-5 synthesis finds agent threats novel but cybersecurity principles adaptable
Published May 18, 2026; summarizes CAISI RFI NIST-2025-0035 (comments due March 9, 2026).
Copilot agents split on declarative vs custom engine before Control System gates
Microsoft docs: declarative agents inherit M365 compliance; custom engine agents bring your own orchestrator and hosting bill.
NIST CAISI opens an AI Agent Standards Initiative on three pillars
The 17 February 2026 launch ties industry-led protocols to agent security research. An RFI on agent security closed 9 March.
NIST CAISI agent-security RFI closed March 9; summary analysis landed in May
Docket NIST-2025-0035 opened 12 January 2026. NIST Trustworthy and Responsible AI 800-5 summarizes responses published 18 May 2026.