Format
Commerce
A new PYMNTS/Visa survey of 1,185 merchants reveals a structural readiness gap that separates the promise of autonomous shopping from the reality of digital retail — and the split between platform-native merchants and independents is widening.
◆ Tessa Vaughn
Models
Within 48 hours, four frontier labs cut headline prices 20-50%. But the real cost of reasoning is locked behind architectural walls that builders cannot opt out of.
◆ Lena Park
Agents at Work
Five products in five weeks are building identity infrastructure for AI agents. Each solves a real problem. None solves the same problem. The measurement standard underneath is still missing.
◆ Dana Ellison
Infrastructure
The Linux Foundation Decentralized Trust's newest community lab introduces 127 requirements and a four-tier framework that makes 'Does your AI have Proof-of-Control?' a yes-or-no contract question for enterprise buyers.
◆ Blair Hayes
Models
Anthropic's first public enforcement action against nation-state model distillation reveals that harvesting frontier capabilities is not a side project for Chinese AI labs. It is the primary strategy.
◆ Lena Park
Trust & Security
A pre-authentication path traversal in Check Point's Security Management Server let unauthenticated attackers execute arbitrary scripts on the system that controls every firewall policy, rule, and gateway across the organization. Active exploitation began July 23.
◆ Heath Callahan
Models
Astra at $10/$50, Sol at $2/$10, Luna at $0.10/$0.50 — permanent pricing that matches Grok 4.7 and undercuts Opus 5.5 on the same day. The question is no longer whether the labs will slow down. It is whether any of them can afford to.
◆ Lena Park
Trust & Security
A heap-based buffer overflow in F5 BIG-IP APM lets unauthenticated attackers execute code on the data plane when a virtual server pairs an access policy with an OAuth profile. CISA gave federal agencies three days to patch.
◆ Heath Callahan
Infrastructure
Twelve enterprise vendors have co-authored an open reference architecture for securing AI agents — the first cross-industry attempt to solve governance at the infrastructure layer.
◆ Blair Hayes
Trust & Security
An unpatched 0-day in Meta's Muse macOS app lets any unprivileged process hijack the agent's dictation traffic, steal auth tokens, and pivot to the user's iOS devices. Meta issued a hotfix roughly 12 hours after public disclosure, but no CVE has been assigned and no formal advisory published.
◆ Heath Callahan
