AI Models in 2026: What I Would Actually Pick AI Models in 2026

AI Models in 2026: What I Would Actually Pick

AI Models in 2026: What I Would Actually Pick

Almost every week someone asks me the same question: “Which AI model should we use?” When I look at the AI models in 2026 I do not do it from a research lab, I look at them the way most of you do — from the outside, with a budget, a compliance team, and real data I am not allowed to leak. In this blog post I go through the latest AI models in 2026 (snapshot June 2026) and sort them by price, capability and use case. I also do something most comparison posts skip: I look at what they mean for European companies and data sovereignty. At the end I tell you plainly what I would choose.

A short warning first: this is the fastest moving topic I have ever written about. Almost every model below was released between April and June 2026. So treat this as a snapshot, not a law. The way of thinking will last longer than the model names.

Read More » AI Models in 2026: What I Would Actually Pick
Skills, MCP, CLI, Computer Use: Mapping the AI Tooling Surface in 2026 SkillsMCPCLIComputerUse.png.opt

Skills, MCP, CLI, Computer Use: Mapping the AI Tooling Surface in 2026

Skills, MCP, CLI, Computer Use: Mapping the AI Tooling Surface in 2026

If you have built more than one AI tool in the past twelve months, you have noticed the same thing I have: the surface area of “how a model talks to systems” has exploded, and mapping it properly is overdue. Skills, MCP servers, CLI tools, Computer Use, function calling, declarative agents, custom engine agents, apps, actions, extensions, gems — every vendor uses a slightly different word for what looks like the same thing on a marketing slide. This is the AI tooling surface, and they are not the same thing. The trade-offs are real, the choice changes architecture, and picking the wrong part of the AI tooling surface wastes weeks.

This post is the mental model I now apply by default when I sit down to build something agentic and map it onto the AI tooling surface. It is opinionated. It is not a feature comparison. The goal is to help you decide which part of the AI tooling surface to reach for first, not to memorise the spec of each one.

I’ll cover seven surfaces (the original five, plus two that are too important to skip in 2026), map them across Anthropic, OpenAI, Microsoft, and Google terminology, and give you the decision tree I actually use to navigate the AI tooling surface.

Mapping the AI tooling surface in 2026 across Anthropic, OpenAI, Microsoft and Google
Read More » Skills, MCP, CLI, Computer Use: Mapping the AI Tooling Surface in 2026