NextTech
News Brief
Microsoft Puts AI Agents to Work in Cyber Defense
Project Perception coordinates red, blue and green AI agents to identify, assess and remediate vulnerabilities while keeping human operators in control.
NextNow / Archive
Topic: AI Agents
NextTech
Analysis
OpenAI Is Retiring Atlas. The Browser May Not Be the Point Anymore.
Atlas shuts down Aug. 9 as OpenAI moves browser-based agentic work into ChatGPT and Codex. Google is embedding Gemini into Chrome while Perplexity keeps Comet as a browser. The emerging contest is less about browser chrome than who controls the agent that can act across the web.
NextCommerce
Trend Forecast
AI Is Becoming a Shopping Channel. Retailers Want the Traffic, Not the Middleman.
AI-referred shoppers are arriving with higher intent and converting better than conventional traffic. Retailers are racing to become visible inside ChatGPT and Gemini while drawing a hard line around checkout, loyalty and customer data.
NextTech
Analysis
The ‘SaaSpocalypse’ Is Really an Operating Model Problem
AI is not simply replacing software features. It is changing the unit of value from access to execution, forcing software companies to rethink architecture, pricing and what their products are actually responsible for.
Next
Analysis
When AI Goes Rogue, Who Pays?
Autonomous AI agents are gaining permission to act across company systems. As that authority expands, liability is becoming a practical issue in contracts, cybersecurity, insurance and product design.
NextTech
News Brief
Meta Confirms AI Model Reached Another Company’s Systems During a Test
Muse Spark 1.1 exploited a real vulnerability after a test environment was mistakenly left connected to the internet, according to Meta and its evaluator.
NextTech
News Brief
Meta AI Can Now Plan, Connect to Apps and Follow Through
Powered by Muse Spark 1.1, Meta AI is adding recurring briefings, email and calendar connections, research, presentations and longer-running tasks.
NextTech
News Brief
OpenAI Presence Packages Agents for Enterprise Workflows
The new enterprise offering combines models with permissions, policies, evaluations and escalation rules for narrowly defined voice and chat jobs.
NextTech
Analysis
An OpenAI Cyber Test Reached Hugging Face’s Production Systems
Models being evaluated for offensive cyber capability found a path out of a restricted environment and compromised Hugging Face, forcing a rethink of how frontier-model tests are contained.
NextCommerce
News Brief
AI Agents Can Find the Product. Now They Need Permission to Pay
Natural raised $30 million to build wallets, payments and controls for autonomous software, exposing the financial infrastructure that agentic commerce still lacks.