
By Geethika Isuru ·
You're Already Building in Public. You Just Don't Know It Yet.
It is 2 in the morning. You are in front of Cursor. You hold the mic key.
Okay Cursor I think the bug is in the webhook handler. Stripe is sending the event. We are receiving it. But the user record is not updating. Let me check the logs. Wait, I see it. We are filtering by email and the email is null on this account because they signed up with Google. Fix the filter. Use the Stripe customer id.
You release the key. Cursor writes the code. The bug is fixed.
Then you close the laptop and go to bed.
You did not realize it but you just gave away three LinkedIn posts. None of them got written. None of them got published. The story is still in your head and now it is gone.
This happens every day to every vibe coder I know. Including me.
The disconnect
Here is the math. You voice typed for two hours today. Two hours of you explaining decisions, debugging, talking through a feature.
You posted nothing.
The 2 hours of voice contains every ingredient of a build in public post. The decision. The failure. The fix. The lesson. The exact words you would have used if you sat down to write later. They are already there. You said them. They went into a prompt and disappeared.
You are not failing at build in public. You are doing it. You are just deleting the recording.
What you actually say to Cursor
Watch yourself for one full session. The sentences you say into Cursor break down into a few patterns.
- The plan. "I want to add a feature that does X. The user flow is A then B then C."
- The decision. "Use approach 1, not approach 2, because the second one creates a circular dependency."
- The debug. "It is failing. I think it is the database. Let me check. Yeah, the migration did not run."
- The reaction. "Whoa wait. That is way faster than I expected."
- The reflection. "I should have done this on day one. We have been losing data for two months."
Every one of those is a post. The plan is a build update. The decision is a story. The debug is a war story. The reaction is a clip. The reflection is the kind of post that goes viral because it sounds human.
You already make all of them. Daily. Without trying.
A real example
Here is a scene from a vibe coder I know. He was working on a Stripe integration.
The voice he gave to Cursor went like this.
I am going to handle the subscription cancelled webhook. When that fires, mark the user as free in our database. But also, do not delete their data. They might come back. Just flip the plan field. Wait, also I need to handle the case where they had pro features active. Like custom domains. Do we shut those off immediately? Probably not. Probably give them a grace period. Seven days. Then run a cron to clean up. Yeah that feels right.
Six sentences. Maybe forty seconds of voice.
That same vibe coder later posted this on LinkedIn.
Spent today thinking about churn. Specifically, what happens the second a customer cancels. The default move is to flip them to free and shut everything off. That is the cold version. The retention version says wait. Give them seven days. Custom domains, integrations, everything still works. Why? Because if they come back, they should not have to set things up twice. Friction kills returns. Tiny detail. Big impact on win backs.
Same content. Different shape. The post took zero new thinking. The thinking already happened during the build.
Three types of moments hiding in your sessions
You have three goldmines in every voice session.
- The decision moment. You explain why approach A beats approach B. That is gold. People love decisions with a reason behind them.
- The failure moment. Something breaks. You debug out loud. You say what you tried, what failed, what you are about to try next. Add the resolution and you have a thread.
- The user moment. Slack pings. A user said something. You react out loud. "Three people in a row asked for export. We need to build that." That is the kind of post that converts because it is real.
You make all three every week. The question is just whether you keep them.
Why no other tool can do this
Here is the moat that nobody talks about.
Generic AI post generators ask you to type a topic. They invent a voice. The output is fine. It also reads like every other AI post on LinkedIn. The reader scrolls past.
Mahasen and tools like it start from your voice history. The words in the post are mostly your words. The tool just removes the filler and shapes the arc. The AI is editing, not pretending.
That is why no competitor with a "topic to post" workflow can produce this. They do not have your voice. They have a blank prompt and a model. The thing they generate is generic by construction.
Your voice is the moat. The fact that you talk while you build is the unfair advantage.
How to capture it without effort
You have three options. Pick the one that fits your stack.
- Use a voice typing tool that already keeps a session history. You voice type into Cursor like normal. The history saves on your machine. You pull from it weekly to make posts.
- Use voice notes on your phone. Talk through your work. Save to a notes app. Same idea, lower friction, less context.
- Use a tool like Mahasen that does the capture and the post drafting in one loop. You voice type. It saves the history. It surfaces story moments. You hit copy and post.
The tool matters less than the principle. Stop letting your voice disappear into prompt windows. Save it. Then mine it.
What to do this week
Pick a single dev session. Voice type the whole thing. At the end of the session, scroll back through what you said. Find three moments. Write three posts.
If you cannot find three, try again tomorrow. Most days you will find five.
The point is not to post every one. The point is to prove to yourself that the content is already there. You just need to stop deleting it.
You are already building in public. The only choice you have left is whether you let anyone see it.
See your voice session become posts
Mahasen captures your voice typing while you build, then turns it into LinkedIn, X, and Reddit posts that sound like you. The dev session is the input.
Frequently asked questions
How do I turn voice typing into LinkedIn posts in May 2026?
Capture the voice history first. Most voice typing tools throw the audio away after the transcription. The tools that work for build in public keep the session and let you mine it. In May 2026 the cleanest workflow is voice type into Cursor, save to a tool that holds the history, then ask an AI to surface the post moments.
Can I build in public from my Cursor sessions in May 2026?
Yes, and this is the easiest way to do it now. Every Cursor session contains your decisions, your debugs, and your reactions. Those are the three building blocks of a good build in public post. Tools that connect to your voice history make this almost zero effort. Mahasen is one of them.
What should I post when I am building in public as a vibe coder?
Post the decision, the failure, and the user moment. Skip feature updates. Skip vague status posts. The vibe coder posts that work in May 2026 read like a story. They have a hook, a turn, a landing. The good news is you already make those stories every time you talk to Cursor.
Is voice typing actually faster than typing for vibe coders?
For most people yes. Voice typing runs at 150 plus words per minute. Most developers type at 60 to 80. The bigger win is not raw speed. It is that voice surfaces context you would have skipped if you were typing. You explain more out loud. That extra context becomes both better code and better content.
How do I make my voice into authentic LinkedIn posts?
Start from your real words. Most AI tools start from a topic and invent a voice around it. The fix is to feed the AI your real sentences from a voice session. The output reads like you because most of it came from you. The AI is editing, not pretending.
What is the best AI for turning voice notes into posts in May 2026?
The best one in May 2026 is the one that starts from your voice, not from a topic prompt. Tools like Mahasen capture your voice history and turn it into LinkedIn, X, and Reddit posts. Generic AI writers will not work because they do not have your context. The voice is the input, not the topic.
Can vibe coders build an audience without writing posts manually?
Yes, and this is the new pattern in May 2026. Vibe coders are using their voice typing history as the source. They voice type to Cursor and Claude every day. A tool then turns that history into posts. The dev session is the marketing input. Nothing extra to write.
How long does it take to make a post from a voice session in May 2026?
Under thirty seconds with the right tool. You finish a build session. You open the tool. You see three story moments pulled from your voice. You pick one. You hit copy. You paste into LinkedIn. Done. The whole loop is faster than reading a typical AI generated draft.
From the Mahasen field notes on voice typing for founders.