Hi there!

Welcome to the 33rd edition of Work in Beta.

In this edition, we demystify markdown files: what a .md file actually is, how to read, write, and view one, and why AI made this format show up everywhere.

So, let’s dive in!

In a recent workshop, a participant told us: "The moment these markdown files appear, I feel very nervous. I don't know how to write these files or work with them."

An AI tool hands you a file ending in .md. You open it: bare text, # marks and asterisks scattered through it. You can read the words, but you don't know what the marks mean, whether you're allowed to change anything, or how this plain text is supposed to become a proper document.

Here is the reassurance: a markdown file is a plain text file with labels on it.

By the end, you'll have written one yourself and know how to view any of them properly.

Here is a complete markdown file:

```

# Weekly update

Client work is on track.

- Send the proposal to Meera

- Book the review call for Friday

**Deadline: 22 August * *

```

Type those lines into any text editor, save it as notes.md, and you have written one.

The labels are the only new part and highlight how the text is formatted and signal hierarchy. For example, A “#“at the start of a line means "this is a heading". A “-“ means "this is a list item". Two asterisks around words mean "make this bold". Those three cover most markdown files you'll ever open.

And it's not new. A writer named John Gruber created markdown in 2004 to add headings and bold to his writing without making the file ugly to read. That's the entire idea: the file stays readable.

One question remains: the raw file is readable, but how do you see the nicely formatted version?

Two easy ways, nothing to install:

- Open it in Google Docs. Google Docs works with markdown. Open your .md file there (File, then Open) and it becomes a normal formatted document.

- Use a free preview site. StackEdit is one: paste your text on one side and the formatted page appears alongside.

The technical crowd does this in code editors like VS Code; if that sounds daunting, the two options above are enough.

That's how you see these files. What the AI sees is a different story, and it explains why markdown is everywhere now.

When you upload a file to an AI tool, the AI doesn't open it the way you do. A converter pulls the text out first and hands it over. The AI works on whatever survives that step.

Some things don't survive. OpenAI's documentation says charts and images inside Word and PowerPoint files don't reach the model. Google says the same about Gemini: those files come through as plain text only. So the slide where one chart carried your whole argument arrives as sentences with the chart missing, and the AI summarises what's left without telling you.

That's not a reason to stop giving slides to AI, just to share them the right way. When the charts matter, save the deck as a PDF first: most AI tools send along a picture of each PDF page, so the model actually sees what you see.

And a markdown file? It has nothing to lose in that step. It's already plain text, so what you see is exactly what the AI gets.

Hold on to that fact: AI reads and writes plain text; everything else is converted on the way in. A format that already is plain text gives the converter nothing to break.

You see the proof every day. Ask ChatGPT or Claude any question and look at the shape of the answer: headings, bullets, bold phrases. That formatting is markdown. The AI writes its answers with the same labels you just learned; the app displays them nicely.

AI tools also hand these files to you now: exported notes, project files, and the CLAUDE.md and AGENTS.md instruction files we broke down in May. Every edition of this newsletter, this one included, starts as a markdown file on our machine.

The part we care most about: you and the AI are reading the same file. The # the AI sees is the # you see. So when the AI keeps getting something wrong, you can open the very file it reads and fix the line yourself. That's why the style guide we had you build in April is a markdown file: correct it today and your AI works from the corrected version next session.

Not every file should become markdown. People mix up two questions here: what format a file should live in for your work, and what you hand to AI. The table answers both.

Small tables are fine inside markdown; the one above is markdown. The moment you want to sort a table or total a column, it's data, and it belongs in a CSV file or a spreadsheet.

.docx and .pptx never appear in the last column. Everything Word and PowerPoint add on top of plain text (comments, tracked changes, slide design) exists for people, and the converter strips it out before the AI reads a word. So handing AI a .docx gains you nothing over the plain text inside it. That doesn't mean abandoning Word: if your report lives there, keep it there. When AI needs it, paste the text or convert to markdown.

A second reason the last column leans toward plain text: AI reads and writes in tokens (small pieces of text); every extra token spends time and usage allowance. Cloudflare measured one of its own web pages, and the AI needed five times as many tokens for the web-page version as the markdown version.

A note on the other direction, when AI produces the document. Ask ChatGPT or Claude for a presentation and it doesn't open PowerPoint: it writes plain text, and a tool packages that into the .pptx or .pdf you download. So draft in markdown or plain chat text, where revisions are cheap and you can edit any line, and ask for the finished file once the content has stopped changing.

That participant in our workshop wasn't missing any technical skill. They were looking at plain text with a few labels on it, and nobody had ever told them that's all it was.

So here's our ask for the week: write one markdown file. Your notes from one meeting, or the instructions you keep retyping into AI. One # title, a few - bullets, saved as notes.md. Then open it in Google Docs and watch it become a proper document. The nervousness doesn't usually survive the first file.

Most professionals know AI can do more for them. The gap isn’t awareness - it’s knowing where to start, what to change, and how to make it stick.

That’s what we work on through Work in Beta.

For individuals, we run working sessions, not teaching sessions. You bring a real problem from your actual work; we build the solution with you, live. You’re borrowing our learning curve instead of grinding through your own. You walk out with something that works and the muscle to keep going. When you get stuck later, we’re a message away.

You probably have a version of at least one of these:

Bring that. We’ll build it with you.

For organizations, AI adoption is a people problem, not a technology problem. Your teams have the tools, what’s missing is the translation layer between AI capability and daily work. Which processes to redesign, which habits to break, how to build genuine fluency, not just awareness. We help close that gap through hands-on training, process redesign, and deep adoption engagements. Not advisory, forward-deployed.

If any of this resonates, email us at [email protected] and we will figure out how to work together.