Welcome Back to The Cwtch 🧡
A cozy corner of the internet where AI makes sense — no degree required.
This week: AI is moving from answering questions to taking action. That sounds exciting. It also means your boundaries matter more than ever.
🤖 THIS WEEK'S STORY: AI IS BECOMING LESS LIKE A CHATBOX
For the last couple of years, most people have experienced AI as a chat window.
You ask a question. It answers.
You ask for a rewrite. It rewrites.
You ask for ideas. It gives you a list.
That version of AI is already useful. But the next version is different.
The big companies are now racing toward AI agents — tools that do not just answer you, but can actually help carry out tasks.
That might mean:
searching across documents
comparing options
booking or scheduling things
managing inboxes
building spreadsheets
drafting replies
using software on your behalf
In plain English: AI is moving from talking about work to touching the work.
That is a big shift.
And for regular people, the important question is not, "Which company has the smartest model?"
The better question is:
What am I comfortable letting AI do?
Because once AI can take action, trust becomes practical. Not philosophical. Practical.
You do not need to understand every technical detail. But you do need to know the difference between:
AI that suggests an email
AI that sends the email
AI that sends the email to the wrong person
Those are not the same thing.
The future of AI is not just about intelligence. It is about permission.
⚡ QUICK WIN: THE THREE-DOOR RULE
Before letting any AI tool act for you, sort the task into one of three doors:
Door 1: Suggest only
AI can help you think, draft, summarize, compare, or explain.
Examples:
"Draft a reply to this email."
"Summarize this document."
"Give me three options for this decision."
This is the safest door. Nothing happens until you choose.
Door 2: Prepare, but ask before acting
AI can do the setup, but it needs your approval before anything leaves your hands.
Examples:
"Prepare a calendar invite, but do not send it yet."
"Write the message, then wait for my approval."
"Find the best option, but let me decide before booking."
This is where most people should spend a lot of time.
Door 3: Act automatically
AI can take action without checking first.
Examples:
auto-send routine emails
auto-file receipts
auto-sort documents
auto-run a recurring report
This can be useful. But it should be reserved for boring, repeatable, low-risk tasks.
If the task can embarrass you, cost money, expose private information, or affect another person, it probably does not belong behind Door 3.
Simple rule:
If you would be nervous letting a new employee do it unsupervised, do not let AI do it unsupervised either.
🌍 CWTCH WATCH
The agent race is heating up.
Google recently pushed more tools for building AI agents. OpenAI and Anthropic are also moving deeper into agent-style workflows — not just better answers, but software that can help perform specialized tasks.
That sounds like a business story, but it matters at home too.
Because the more capable AI becomes, the more ordinary people will be asked to trust systems they cannot fully see.
Watch for these three shifts:
From chat to action: AI will increasingly offer to do the next step, not just explain it.
From one tool to many tools: The AI that helps you may touch email, calendars, files, browsers, spreadsheets, and apps.
From convenience to control: The easiest tool may not be the safest one. The safest tool may not be the flashiest one.
The winning question is not, "Can AI do this?"
It is:
Should it do this without me?
That question will matter more every month.
💡 THIS WEEK'S PICK: MANUS
Manus is an AI agent tool built around doing multi-step work, not just answering questions.
Instead of only asking it for a paragraph or a list, you can give it a goal and let it plan the steps, gather information, organize the result, and hand you something more complete.
Good for:
researching a topic and turning it into a simple report
comparing options before a decision
organizing messy information into a plan
testing what agent-style AI feels like without needing to build anything yourself
How to try it:
Go to manus.im and give it a small, practical task:
"Research three affordable weekend trip ideas within two hours of Moncton, compare cost, driving time, and best fit for a quiet getaway."
Then watch how it breaks the work into steps.
That is the useful part this week: Manus gives you a feel for where AI is heading. Not just chat. Not just answers. More like a helper that can take a goal and work through the messy middle.
Still check the result. Still use your judgment. But if you want to understand AI agents, Manus is a good place to start.
ONE LAST THING
AI agents are not something to fear.
But they are something to design boundaries around.
The goal is not to let AI run your life. The goal is to make it useful enough to reduce the boring friction, while keeping you in charge of the things that matter.
Start with Door 1. Get comfortable with Door 2. Be very picky with Door 3.
That is how you get the benefit without handing over the steering wheel.
Hit reply: what is one task you would let AI prepare for you — but not do automatically yet? 🧡
The Cwtch is published weekly. Unsubscribe anytime — no hard feelings.
This newsletter may contain affiliate links. They cost you nothing.