AI Agents News - AI Updates

AI-generated conceptual editorial illustration of a digital workflow passing through a controlled gate with a separate return receipt, without readable text

Marinela Profi discusses agent governance in AI News. Our editorial assessment explores three hypothetical tests for approval, retries and cancellation, rather than treating a demonstration as deployment evidence.

Minimal editorial illustration of a small luminous courier-like form approaching a translucent gate, representing permission-based access for an AI agent

View Post

AI Agents

Website Permissions Are Becoming a Bottleneck for Consumer AI Agents

Retailers and airlines describe different rules for AI assistants. TechCrunch's reporting separates intentional restrictions from technical failures and early standards work.

Abstract illustration of separate geometric units following independent paths into a map-like network

View Post

AI Agents

Researchers Observe Parallel AI Agents Querying Amap Without Evidence of a Swarm

Preliminary traffic observations suggest multiple agents on Tencent infrastructure querying Amap. The report does not establish operator identity, coordination or an attack.

An autonomous shopping agent stopped at a guarded digital marketplace entrance

View Post

AI Agents

Amazon Blocks Meta’s Muse From Shopping on Its Store

Amazon has rejected shopping requests from Meta's Muse agent, exposing how platform permission, liability and model errors can limit agentic commerce.

Specialized AI agents coordinating freight, inventory and warehouse operations across a luminous supply network

View Post

AI Agents

Multi-Agent AI Moves Supply Chains From Advice to Execution

Enterprise trials show specialized agents acting on logistics data, while strict spending limits and approval gates define where autonomy can safely operate.

A supply-chain control room where an AI agent moves a delayed shipment from alert to authorized action

View Post

AI Agents

Supply Chain AI Still Sees Faster Than It Acts

An AI News analysis argues that supply-chain systems detect disruption quickly but need bounded authority, transaction access and clearer accountability to respond.

A consumer approves an AI action inside an isolated virtual machine protected by a security filter

View Post

AI Agents

Meta’s Muse Agent Puts Consumer Trust Inside the Architecture

Muse combines broad consumer actions with isolated virtual machines, external monitoring and a planned confidential-computing layer, while important trust limits remain.

An autonomous coding agent continues a chain of tasks across several sessions while a human retains a clear stop control

View Post

AI Agents

OpenAI Tests a Codex Agent That Keeps Working Until Stopped

Public code shows OpenAI testing a persistent Codex mode that can create follow-up tasks, continue across sessions and message users proactively.

A small-business owner supervises an AI agent connecting a finished digital storefront to customer growth channels

View Post

AI Agents

Runable Raises $21 Million to Move AI Agents From Building to Growth

Runable wants its small-business agent to progress from creating sites to finding customers, while acknowledging that subsidized AI usage leaves gross margins negative.

A continuous assistant memory represented by luminous threads connecting conversation and project workspaces

View Post

AI Agents

Claude Unifies Memory Across Chat and Cowork

Anthropic is giving Claude chat and Cowork a shared, user-editable memory layer, with separate controls for sensitive information.

Follow AI Updates

AI Agents

SAS Interview Puts the Lifetime of Agent Permissions Under Scrutiny

Website Permissions Are Becoming a Bottleneck for Consumer AI Agents

Researchers Observe Parallel AI Agents Querying Amap Without Evidence of a Swarm

Amazon Blocks Meta’s Muse From Shopping on Its Store

Multi-Agent AI Moves Supply Chains From Advice to Execution

Supply Chain AI Still Sees Faster Than It Acts

Meta’s Muse Agent Puts Consumer Trust Inside the Architecture

OpenAI Tests a Codex Agent That Keeps Working Until Stopped

Runable Raises $21 Million to Move AI Agents From Building to Growth

Claude Unifies Memory Across Chat and Cowork

AI Incident Cleanup Needs More Than a Technical Explanation

Anthropic’s METR Access Opens a New Test of Independent Scrutiny

UPS’s AI Parts Search Faces a Test of Evidence and Eligibility

Singapore’s On-Device Scam Detector Separates Privacy From Proof of Protection

Ricursive’s Chip-Design Strategy Extends Beyond Faster Floorplanning

Flock’s Published Audio Policy Reaches Beyond Gunfire

A Theory-Research Study Measures AI Progress Without Declaring Science Solved

Zenzap’s Funding Brings Shared Workplace Agents Into View

Stay Ahead of AI

Follow AI Updates