Manuviraj is a tech enthusiast who covers Windows, macOS, and retrocomputing. Drawn to both the logic and the creativity of technology, he’s especially interested in how people interact with tech and the internet every day.
He holds an English Literature degree. His fascination with tech grew alongside a love for speculative worlds: cyberpunk and sci-fi literature shaped the way he thinks about the future of machines and humans sharing space.
Outside the screen, Manuviraj’s world runs on two wheels and good sound. He’s into motorcycling, constantly exploring new music, and tinkering with digital tools just to see what they’re capable of. Whether he’s breaking down complex software topics or diving into niche corners of tech culture, he’s driven by the same thing: figuring out how things work, and how they could work better. You can also read his work on his Substack.
It wasn't before I switched to Obsidian as my note-taking app that I began using Markdown. I'd heard of it before because one of my favorite hacktivists, Aaron Swartz, significantly contributed to its creation in 2004. But Markdown gained significant mainstream adoption after GitHub launched in 2008 and used it for README files. Chances are you've already used it without realizing it . However, it's become even more important now, thanks to the proliferation of LLMs and AI agents.
Using Markdown elicits better responses from your AI tool
While using less tokens
There has always been a clear differentiation between machine-readable content and human-readable content. Take HTML for instance: reading raw HTML code is unintuitive, but a browser reads it and renders a beautiful website for me to see. There's a huge disconnect between what I see and what the browser renders. The difference is less noticeable in Markdown because the actual content isn't buried between complicated tags, brackets, and similar ancillary stuff. The syntax is minimal and uses simple characters: even if you see Markdown files in a plain text reader, you will understand it. A Markdown reader will make it look pretty; that's about it.
Most of us prompt LLMs in a conversational tone, writing down paragraph after paragraph, completely disregarding any structure. Now, this is okay for simple everyday tasks. But if you want your LLM to give you better output, use Markdown. An unstructured wall of text makes the LLM work harder to understand the task, and the provided context. But if you use Markdown, it's a win-win situation: you end up using fewer tokens while getting better outputs in the process. The results are even more noticeable when you're vibe coding simple apps: adding a blob of text will get you very different results than attaching a structured Markdown document that has the stack, features, design, and do-nots listed separately.
Markdown is even more crucial when using AI agents that are executing a set of tasks, accessing different apps, and cross-checking. The instructions that govern AI agents' behavior are often written in plain Markdown files. GitHub Copilot reads a copilot-instructions.md file sitting in your repository, while Claude and ChatGPT's custom instructions field supports and responds to Markdown. Knowing Markdown will let you read, audit, and customize your AI tool's behavior to make it work the way you want it to.
Markdown is everywhere
You can add better formatting in tools you use every day
Reddit, Obsidian, Notion, Discord, WhatsApp, Slack, and innumerable other tools support Markdown. All of them may not support the full gamut of Markdown elements, but they support simplified versions of Markdown. Maybe you've used _word_ and *word* to italicize and bold specific words on WhatsApp already. Or maybe you begin unordered lists in your Slack messages by using a hyphen (-). Yes, that's Markdown too.
I began using Markdown by familiarizing myself with its six most common elements from its syntax, and I haven't actually felt the need to go deeper:
-
Headings:
# Heading,## Heading,### Heading, for H1, H2, and H3 respectively. -
Bold:
**word** -
Italics:
*word* -
Bullet list:
- item -
Link:
[text](URL) -
Blockquote:
> text
You can further familiarize yourself with the Markdown syntax using this cheat sheet on Markdown Guide.
Coming back to Obsidian for a bit. One big reason for its wide appeal is that your notes are accessible forever, even if Obsidian disappears tomorrow. You have Markdown to thank for this: Obsidian notes are saved as Markdown files (switch to Source View to see the syntax), and Markdown itself is a free and open-specification language. You can open .md files in Notepad, Word, TextEdit, or practically any text editor on any device; it will still be legible. Unlike apps that use closed source formatting, Obsidian, because it uses Markdown, can ensure your notes remain portable and secure forever.
- OS
- Windows, macOS, Linux, Android, iOS, iPadOS
- Developer
- Dynalist Inc.
- Pricing model
- Free
- Initial release
- March 30, 2020
Obsidian is a local-first, Markdown-based note-taking application that stores your notes as plain text files and lets you build interlinked “vaults” of knowledge. It supports plug-ins, graph visualisations, and full control of your data rather than locking you into a proprietary format.
Markdown is easy to learn
If I'm being honest, it'll take you about 20 minutes of deliberate effort to get a grip on Markdown's basics. Since it was designed as shorthand for basic HTML markup, it's minimal, and the six main elements I listed above are a solid starting point. But you can go deeper and learn how to add images to Markdown docs, assign variables to the main elements of Markdown and more. Maybe try prompting your favorite AI tool exclusively in Markdown, and then explore the rest of its capabilities.