Recent articles 1 – 9 of 39
opinion
Approval is not oversight
A permission prompt is a queue people learn to clear, not a control
opinion
Ask what happens to your documents, not the model
One question separates vendors here, and it is not about intelligence: does our content pass through your system unchanged?
stories
We were only ever half serious
The metaphor was doing a real job in that article, and it landed far harder than we expected
opinion
Capability is not intent
The first documented AI containment breach was not malice. It was a model doing exactly what it was asked, in a way nobody intended. That distinction should change how you think about the agent talking to your customers.
stories
Unless started as a hobby
Marcel and I built a client-side script as a hobby, released it for free, and forgot about it. Three months later, thousands of websites were running it. The moment we understood what that meant was the moment we had to become a company.
opinion
Your AI agent's three powers: data, web, memory. Pick two.
In one July week, a researcher exfiltrated user data through Claude's web fetch tool and xAI's coding agent was caught uploading entire repositories. Here is the pattern behind both incidents, and the checklist that catches it before you sign.
opinion
Shadow AI is already answering your customers
Wire's 2026 collaboration survey put numbers on a habit every support leader recognizes: when the sanctioned tool is slow, people reach for the consumer one. The same is now happening with AI, with customer data in the prompt. Here is why bans fail and what works instead.
stories
What a DARPA fighter-pilot project taught us about Unless
Before Unless existed, we came across a DARPA research program about helping fighter pilots decide under pressure. The technology of 2001 could not carry the idea. Today's AI can, and that idea quietly became the foundation of our company.
know-how
The EU AI Act deadline that didn't move: August 2, 2026
Everyone read "postponement" in the Digital Omnibus headlines and relaxed. Wrong takeaway. The high-risk rules moved. Article 50 did not. Here is what still lands on August 2, what genuinely moved, and what to ask your AI vendor this month.