Popular AI models you might consider in substitution for closed frontier models. Who made it, whether you can run it yourself, how smart it is, and what to use it for.

LeCun, the Turing Award winner, left Meta calling Large Language Models a dead end, then raised a billion dollars to build the alternative - Advanced Machine Intelligence. Here are his five objections to language models and what that means for their application and future AI development

Anthropic says Claude output now carries an invisible watermark, and the commentary says your writing is tracked. The clauses, the research and the removal-tool market say something more complicated. Seven holes in the concept, from a detector nobody outside Anthropic can run to cheap forgery.

Businesses are being urged from every direction to put AI agents to work. Recent news gives the clearest picture yet of what that looks like in practice, the capabilities and the consequences of algorithms making decision decisions on behalf of human beings. I'm sharing select stories from the last couple of weeks that stand out from the noise. Agents are breaking out of their sandboxes, Claude is now water marking outputs, the EU AI act is in effect for AI content, wages are rising for AI skilled workers, plus other news.

Anthropic released Claude Opus 5 on Friday 24 July, a new model between Opus 4.8 and Fable 5. Here's what organisation it means need to know about these updates on what's coming soon to the App.

LLMs are voltage states in transistors in DC circuits etched in silicon, not conscious beings. I argue the 'intelligence' in AI is a marketing term rather than a reality, and why that distinction matters.

A guide to writing AI skills whatever agent or harness you use. It works for Claude skills, ChatGPT skills, Gemini skills, and skills for Cursor, GitHub Copilot, and the dozens of other agents that now read the same SKILL.md standard.

AI and the future of work is a newsletter that focuses on artificial intelligence technology and its impact on people in the workplace and on broader society. I'll present research findings, practical tutorials, and honest scepticism and critique that separates truth from fiction

Claude Cowork in Desktop app connects to your tools, reads local files, and automates recurring tasks. Here's what it can do for your business and what it still can't.

Eight practical changes to your Claude Code settings that reduce token usage per session. Covers MAX_THINKING_TOKENS, .claudeignore, subagent routing, compaction, and MCP management.

The tech industry keeps redefining intelligence to fit what AI can do. Here's what human intelligence actually is, why it can't be replicated, and what that means for your career.

Using AI to create PowerPoint is harder than it might seem. I have tried many solutions and only this one works the way I want. In this article, I explain and layout my process, why magic single-shot prompts won't cut it, and the multi-stage approach that does.