Blog
Practical help for owners buried in follow-ups, quotes, customer questions, and office work.
Anthropic's Memory Push Raises the Bar for Governed AI Employees
Anthropic launched persistent memory for AI agents on April 23, 2026. Rakuten reported 97% error reduction. The real story: memory architecture is now a governance problem, not a plumbing problem.
Perplexity Opens Up Its Personal Computer AI Assistant To All Mac Users
Perplexity just opened its Personal Computer AI agent to all Mac Pro subscribers — down from $200/month. The paywall drop is the headline. The cloud-vs-local architecture is the story that matters.
Amazon quietly builds an agentic commerce team to connect with ChatGPT
Amazon posted a job listing for a Principal TPM of 'Agentic Commerce Experiences' — a 40-person team building APIs for ChatGPT and other AI platforms. It's not a reversal of its anti-scraper stance. It's the next phase of the same strategy.
Amazon Bedrock AgentCore Payments: AI Agents Can Now Pay for APIs and Data
AWS launched AgentCore Payments in preview. AI agents can now pay for APIs, data, and services inside the execution loop. Here's what that changes for governed agent operations.
It's game over for Copilot on Xbox
Microsoft's Xbox CEO just killed Copilot on console — a product that never shipped and was announced at GDC less than two months ago. Here's what the failure reveals about where AI agents actually work.
Securing Your SaaS and Data in the Age of AI Agents
AI agents inherit executive-level access, run 24/7 without MFA, and can be manipulated by content already in your database. Here's what the new agentic attack surface looks like — and how to secure it.
'Infrahub' maker OpsMill raises $14M to give AI agents a single view of
OpsMill raised $14M to give AI agents a single, trustworthy view of enterprise infrastructure. The real story: fragmented data — not AI capability — is what's causing $300K/hr cascading failures.
Meta plans advanced ‘agentic’ AI assistant for consumers
Meta's internal AI agent 'Hatch' is targeting internal testing by June 2026 — and Zuckerberg's own words about existing agent platforms being 'pretty rough' tell you exactly what Meta is trying to solve. Here's the practitioner read.
SoundHound Launches Self-Learning AI Agent Platform
SoundHound launched OASYS on May 5 — a self-learning AI agent platform where agents build and improve themselves. 80% of companies using AI see no ROI. Here's why OASYS is an attempt to fix that structurally.
Kakao readies Kanana 2.5 in push for AI agent platform
Kakao just announced Kanana 2.5 — a model built from scratch for AI agents, not chat. Less than 10% the size of leading models, but engineered to outperform them where it counts: planning and taking action. Here's the real story.
ServiceNow Extends Build Agent to Cursor and Claude Code with Governance Attached
ServiceNow is pushing Build Agent beyond its own tools and into environments like Cursor and Claude Code, while keeping governance, approvals, and audit trails on the platform side.
Big Tech’s Personal AI Agents Are Coming for the to-Do List
Google and Meta are both building personal AI agents — Remy and Hatch — targeting the 3.2M users OpenClaw reached but couldn't keep. Here's what changed, why it matters, and what to do about it now.