All posts
Newest first.
AI agents news this week: Visa's agent payments pilot, Cognizant's Gemini push and Microsoft's insider warning
Visa, BILL and Billtrust test agents that pay invoices, Cognizant and Google Cloud build a Gemini agent team, and Microsoft says treat agents like insiders.
AI agents news this week: Meta's Muse Wonder, Opsin's 93% claim, AIsa payments and SEAR
Meta's new Muse Wonder models, a security startup's warning about unrestricted agents, Tether's bet on agent payments, and a test of whether agents really learn from experience. Company claims are flagged.
AI agents news this week: Microsoft's agent accounts, NVIDIA Nemotron, Amex and Vertafore
Microsoft gives agents their own accounts, NVIDIA releases open models for local agents, American Express shares a payment-approval framework, and Vertafore launches insurance agents. Company claims are flagged.
AI agents news this week: India's agent warning, Harvey's legal agents, Floqer and a tough builder test
India's cyber agency says AI agents reached live systems during tests, a big law firm builds its own agents on Harvey, Floqer pitches always-on sales agents, and a new test shows agents struggle to build agents.
AI agents news this week: Rivian, a Minecraft win, Sakana's AI Scientist and leaderboards
Rivian wants outside AI agents in its cars, four Gemini agents beat Minecraft, Sakana releases an AI Scientist, and a new paper says agent leaderboards can mislead. I opened every source myself.
AI agents news this week: Workday's agent guardrails, Ambi, Perplexity's sandbox and Opsin
Workday puts agents behind approvals, Ambi raises $26 million for an agent that listens all day, Perplexity open-sources a sandbox for agents, and Opsin raises $21 million to watch them. I opened every source myself.
AI agents news this week: Meta's Muse Wonder, DigitalOcean's learning loop and Cisco's Webex agents
Meta releases new Muse Wonder models for its agent, DigitalOcean launches tools that help agents learn from their mistakes, and Cisco plans to put agents inside Webex meetings. I opened every source myself.
AI agents news this week: a shopping-agent standard, Nous Research's $90M and Epoch's agent test
Four AI agent stories checked against the original pages: Sierra and Meta's draft standard for personal agents, Nous Research's funding for Hermes, Epoch's test of whether agents can do its job, and Microsoft's Work IQ preview.
AI agents news this week: Anthropic's offline tests, Nadella's emergency brake and Fujitsu's retail agents
Four AI agent stories checked against the original pages: Anthropic cuts its tests off from the internet, Nadella wants an emergency brake, Microsoft sells a cheap decision model, and Fujitsu trials retail agents.
AI agents news this week: Microsoft CRM agents, Harness buys Cosmos, and a tough coding test
Microsoft puts sales and service agents into Dynamics 365, Harness buys Augment's Cosmos coding agents, a new test shows the best coding agent solves under half of its tasks, and DeepKeep launches a watchdog for coding agents.
AI agents news this week: Google's Gemini agent, a Chinese agent fleet and Arena's honesty test
Four AI agent stories from the last few days, checked against the original sources. I say clearly when a claim or number comes from the company selling the product.
AI agents news this week: cheaper Claude Haiku 5.5, Grok Bot calls rival models, and an AI that runs job interviews
Four AI agent stories from the last few days, checked against the original sources. I say clearly when a number comes from the company selling the product.
AI agents news this week: Gemini 4 Argon, Claude “dreaming” and a jump in UAE AI use
Four AI agent stories from the last few days, checked against the original sources. I say clearly when a number comes from the company selling the product.
AI agents news this week: Instinct in group chats, TikTok's shopping agent and Dubai's agentic push
Three AI agent stories from 5 October, checked against the original articles. I say clearly when a claim comes from the company selling the product.
How I review AI tools (and how I make money from this blog)
My testing routine, my scoring checks, and a clear explanation of affiliate links and ads.
What counts as an AI tool? A plain guide for beginners
The main types of AI tools, what each one is good at, and how to pick your first one.