3 min · every weekday

Three minutes. One prompt, one word, one skill.

Most people feel behind on AI and have no idea where to start. This is the smallest useful thing that fixes that: one email, every weekday, that takes about three minutes and leaves you able to do something you could not do before.

Free. Unsubscribe in one click, any time.

What is in every issue

The same four things, in the same order, so you always know what you are getting and can stop whenever you have had enough.

1 Prompt of the Day
One technique plus a copyable example.
2 Word of the Day
One AI term in plain English plus a one-line example.
3 Try This
One workflow you can run in two minutes that saves time today.
4 Don't Do This
A common AI mistake and the fix.

Then one link at the bottom if you want to go further. One — not four. Most newsletters train you to ignore their calls to action by putting one in every section.

An actual issue

Not a mockup. This is a real issue, built from the same library every issue draws on.

Prompt of the DayStress-test a plan before you commit

You have a plan or proposal and want to find the holes first.

Act as a sharp, skeptical advisor who wants me to succeed. Here is my plan. Do four things: (1) name the 5 biggest risks or weak assumptions; (2) for each, the question a critic will ask; (3) what would have to be true for this to work; (4) the single change that would most improve the odds. Be direct, not polite.

My plan:
[paste plan]

Today's prompt is sponsored by PromptSharp. See PromptSharp

Word of the DayAI Hallucination

A hallucination is when an AI states something false or fabricated — a fake citation, statistic, or quote — with total confidence.

Every word we run becomes a permanent page you can look up later. Browse the glossary.

Try ThisMake AI interview you before it answers

Before asking AI to write or plan anything, tell it: 'Ask me 5 questions that will make your answer sharper before you start.' Answer them, then let it produce the work.

Don't Do ThisDon't ask the model "is this correct?"

Models are agreeable by default. Ask "is this right?" and you will almost always be told yes — including about the parts that are wrong.

Instead: Ask it to attack the work instead: "What are the two weakest assumptions here, and what evidence would disprove them?" You get a critic rather than a cheerleader.

Every mistake we run is collected on one page. Read all of them.