#2. The Weekend Report

The discussions, research and news in AI alignment & governance, since Friday.

1. Freedom of AI Speech

Facebook's shocking, written policies for conversational AI and a man's death.

The worst advice for using AI

PLUS: Making your own WEEKLY FAMILY FUN PLANNER assistant

When an AI "business manager" got told it was AI ... and started demanding to speak to security.

Anthropic tried having an advanced AI Agent run a business. Things were learned.

Get'ya AI superpowers here!

Looking for an easy-button to go past ChatGPT, and do more with AI? Here, we look at one of the fundamentals in approachably more detail; MCP and why you should care.

Set up your foundation for building AI workflows & agents. (Not too technical.)

As always, your weekly guide, under 10 minutes. Plus the news.

3 better ways to write with AI

(without AI writing for you)

How much I'm spending on AI rn.

This spreadsheet was a bit alarming. Where (and how much) I'm spending on AI.

"Trust" is Learned

Thoughts on the tricky trust issue - one we'll need to deal with as AI Agents blossom.

What is the "agent mesh"?

What is Model Context Protocol (MCP)?

What tools do agents have access to, and how do they access?

What is a trigger?

What is the difference between 'deterministic' and 'probabilistic' processes?

What's the difference between AI agents and automation?

How do AI agents work?

Should I think of an AI agent as a replacement for hired roles?

Are there 'types' of AI agents?

How do AI agents differ from chatbots?

What is an AI agent?

Building the hot new AI agent use-case

(At least as far as my kid is concerned) My trip into building the bedtime story agent.

It's not a perfect way to think of personal AI Agents - but close

An insight, a use-case of agentic AI, three new shareable things.