To start before you feel ready, choose a small, useful thing you can finish, decide who will see it and give yourself one specific skill to improve afterward. You can prepare without waiting to feel completely confident. Let the first attempt teach you what the next one needs.
I’m learning this through my podcast.
With each episode, I become a better host. Was I great at the beginning? I don’t know. I assume there were plenty of things I could have done better. Maybe someone listening thought I was fantastic. Maybe they were simply interested in the conversation and weren’t grading my hosting at all.
What I do know is that I’m doing it publicly.
People can hear me learning. They can hear the filler words. The “um” and “uh.” The things I now notice because I’ve actually put myself in the situation.
And sometimes that is fucking hard.
I want to say that clearly because “just start” can sound so easy when somebody else says it. Starting means accepting that the version people see today might be a version you know how to improve tomorrow.
That is uncomfortable. It is also how this is working for me. It’s the same lesson I learned when I quit my job with zero plan: you get clearer once you’re already in the work.
Why is it so hard to start something you actually want to do?
Sometimes starting makes the idea available for other people to judge. Until then, it can stay safely in your head, full of possibility.
You want to launch the podcast. Build the app. Start the business. Share what you know.
But first, apparently, you need a better microphone. Another course. A clearer niche. A more professional website. A little more confidence.
Some preparation is necessary. A real lack of time, money or knowledge can also be a real constraint. I wouldn’t label every obstacle an excuse.
But there is a question worth asking:
If nobody could see my first attempt, would I still need all this preparation?
If you would happily try it privately, the visible part may be what needs your attention.
You can see this in public discussions too. In one small-business thread, a founder described completing the building work and then freezing when it was time to sell. Their concern was putting themselves in front of people. That’s one person’s experience, not proof of what every founder feels, but it captures a very recognisable moment.
The product is ready to meet someone. Suddenly, you find ten more things to fix.
What if you’re afraid of being a beginner in public?
Give yourself a beginner-sized commitment. You don’t have to announce a complete reinvention of your life to try one thing.
A first podcast episode is a conversation you are learning to host. A first app release is a small tool you are learning to make useful. A first offer is something you want to find out whether people need.
You can care about the work without requiring the first attempt to demonstrate every skill you hope to develop.
For me, podcasting is making that distinction very concrete. I want to become a better communicator. Recording episodes is showing me where I want to improve.
I couldn’t have learned all of that by planning the podcast.
There is a limit to what preparation can tell you about how you will handle a real conversation. At some point, you need to have the conversation.
You can practise first. You can prepare questions. You can make the situation manageable. But you still have to let an actual attempt happen.
What does learning in public actually mean?
Learning in public means letting other people see some of your work while you are still developing the skill. You might publish an episode, share a useful lesson, demonstrate a working feature or explain something you recently figured out.
In his essay Learn in Public, Shawn Wang, also known as swyx, encourages people to create and share material from what they are learning. One useful part of that approach is making something that would have helped your earlier self.
You don’t need to position yourself as the final authority on a subject to share an honest observation.
“Here’s what I tried, what happened and what I’d change” is something you can say when that is what you know.
And learning in public doesn’t require broadcasting everything. Keep private information private. Choose what you share and where. You can rehearse privately, test with a small group and publish selected work.
The important step is allowing some contact with people who can actually experience what you’re making.
How do you tell whether you’re preparing or avoiding?
Ask what the next preparation task will change about the attempt.
Checking that the microphone records understandable audio has a clear purpose. Comparing microphones for another week after you already have usable audio may deserve a second look.
Testing that an app saves someone’s work matters. Redesigning the icon again because you don’t want to show the app to anyone is a different situation.
Try writing these two sentences:
Before I share this, I need to finish ______ because otherwise ______.
Once that is done, I will show it to ______ on ______.
If you can’t explain the consequence, the task might be optional. If the finish line keeps moving after you meet it, ask what you are hoping the extra work will protect you from.
Analysis paralysis shows up here too: turning a decision about the next small step into a decision about your entire future.
You don’t need to settle your entire future before you record an episode.
How do you start before you feel ready?
Make the next attempt specific enough that you can do it, review it and learn from it.
1. Choose something you actually care about
There needs to be a reason to come back when the process gets uncomfortable.
What do you want to make? What conversation do you want to have? What problem do you want solved?
If you don’t care about the problem, don’t build an app around it just because somebody else’s results look exciting. Find something you want to spend time understanding.
2. Define the smallest useful attempt
For a podcast, that might be one focused conversation. For an app, it might be one task a person can complete. For a business idea, it might be a small offer you can honestly deliver.
“Small” still needs to mean something to the person receiving it. A podcast needs understandable audio. An app needs to do what you say it does.
3. Decide who will see it
You can start with a relevant, manageable audience. Show the tool to a few people who experience the problem. Ask a trusted listener to comment on the clarity of a short recording. Share a useful post without turning it into a major announcement.
Choose people whose feedback has a connection to what you are trying to learn.
4. Set a finish line for preparation
Decide what must be true before you share it. Write it down while you can still think clearly about the task.
Then set a date that fits the work. When that date arrives, check the agreed requirements instead of inventing a new definition of ready.
5. Pick one improvement for the next attempt
Afterward, ask what worked and choose one specific thing to practise.
“I’m terrible at this” doesn’t tell you what to do next.
“My opening question was too long” does. So does “the user couldn’t find the save button.”
You can do something with those observations.
What podcasting is teaching me about communication
Doing the podcast has made me notice my filler words and think more seriously about how I communicate.
That doesn’t mean I’ve decided every “um” is a disaster. I want to communicate clearly, not turn a conversation into a performance where I’m monitoring every sound I make.
But I now have something specific I want to work on.
And that has given me an idea for an AI tool.
I haven’t built it yet. I want to build a tool that helps me review my speaking and practise the parts I want to improve.
The first version could help me find repeated filler words in a short recording, review the moments around them and choose a small practice exercise. I would also want it to help me notice when a question becomes longer or less clear than it needs to be.
I would need to check how accurately it captures the recording. A transcript can omit the very filler words I want to examine, so a clean-looking transcript wouldn’t automatically tell me how I sounded.
But the point is that I now understand a problem I want to solve because I’m experiencing it.
That is exactly how I think about coming up with app ideas. Frustration, pain, problems and desire give you somewhere useful to start. Be your own first user. If other people experience the same thing, you have something to investigate with them.
The sequence matters: do something, notice the difficulty, then work on a useful improvement.
I don’t need to build the communication tool before I record another episode.
How can AI help you start and improve?
AI can help you prepare a small attempt, rehearse parts of it and organise feedback. Give it a specific job that ends in an action you can take outside the chat.
The following are prompts to try, not claims about tools I’ve already built or results I’ve measured.
Use AI to make the first step manageable
I want to [start a podcast/build an app/test a business idea]. I’m putting it off because [specific concern]. Help me separate practical preparation from optional work. Suggest one useful step I can finish in [available time]. State any assumptions. Don’t give me a long-term plan.
Read the suggestion and decide whether it fits your situation. You are looking for a step you can actually take, not a perfect explanation of why you haven’t taken it yet.
Use AI to rehearse a specific part
I’m preparing to interview someone about [topic]. Here is my opening question. Help me make it shorter and easier to answer while preserving what I want to learn. Then suggest one follow-up I could ask if the answer is vague.
Say the question aloud afterward. A question that reads well on a screen may still feel unnatural when you say it.
Use AI to review an example of your work
Review this short transcript for clarity. Point to one place where a question could be more direct and one thing worth keeping. Use examples from the text. Don’t infer my confidence, tone or speaking pace from a transcript alone.
That keeps the feedback attached to the material you supplied. If you’re practising with a guest conversation, use a short excerpt you have permission to share and remove unnecessary private information.
Use AI to turn feedback into the next attempt
Here is the feedback I received. Separate concrete observations from preferences and vague judgments. Suggest one change to test next time. Don’t treat every comment as something I must implement.
You can use that approach for an episode, a product demonstration or a first customer conversation.
My Building With AI guide covers the broader possibilities for personal tools, workflows and apps. Start with the part that helps your actual work.
Can AI become another way to avoid starting?
Yes. You can keep asking it for a better plan, another critique or a more polished introduction long after you have enough to attempt the thing.
You can become more articulate about your problems in a chat without seeing corresponding change outside it. That raises a useful question for anyone using AI to work through hesitation.
What changed outside the conversation?
Did you record the practice answer? Invite the guest? Show someone the working feature? Ask a potential customer about their current problem?
Give the AI session a finish line. For example: leave with one opening question, then record yourself asking it. Or leave with one short invitation, then send it to the person you intended to contact.
If you keep returning for reassurance, try acting on the smallest useful decision you’ve already made before requesting another plan.
What if you’re afraid to launch your app or business?
Choose a release you can responsibly stand behind and put it in front of the people it is meant to help.
I say ship ugly and fast, but useful. The “useful” part matters.
You can release a simple design. You can leave optional features out. You can say honestly what the first version does and what it doesn’t do yet.
You still need the promised task to work. Protect people’s information and check the parts they will rely on. Being a beginner doesn’t remove responsibility for the product you give someone.
For deciding what belongs in that release, use Simple, Lovable, Complete. For deciding whether people want it, use app idea validation. If you’re building with AI tools, my guide to starting vibe coding covers getting that first small release out before you feel ready.
Then make the learning question concrete: can this person get the result, and do they want to use it again?
A small release gives you something more useful to work with than another month of imagining everybody’s reaction.
What if people judge you or your work isn’t good yet?
Some people may judge it. I don’t think promising otherwise helps.
But you can decide which feedback deserves your attention. A listener saying they couldn’t follow a question gives you something to investigate. A user showing you where a workflow breaks gives you something to fix.
A personal insult doesn’t become useful product or communication feedback because somebody posted it publicly.
Also leave room for the possibility that your assessment of yourself is harsher than the audience’s. I don’t know whether my first podcast episodes sounded as inexperienced to other people as I imagine they might have.
I know I have things I want to improve. I can work on those without inventing an audience that thinks I’m awful.
Frequently asked questions about starting before you’re ready
How do I start something new when I have no confidence?
Choose a small attempt with a clear purpose and limited consequences. Prepare what it actually requires, then do it and review one specific thing afterward. You can make a useful attempt while still feeling uncertain about your ability.
Why do I keep making excuses for something I want to do?
There may be practical constraints, uncertainty about the next step or discomfort about other people’s reactions. Write down what each preparation task will change. If you would try the same thing privately but keep delaying the public version, investigate the visibility concern directly.
How do I overcome perfectionism and actually start?
For this kind of creative or business project, define a specific standard for the first attempt and a date to review it. Include what makes it useful and responsible to share. Once those requirements are met, avoid adding new ones simply because sharing still feels uncomfortable.
Do I have to learn in public to become good at something?
No. Private practice can be valuable, and you don’t have to publish every attempt. But if your goal involves listeners, customers or users, plan some contact with them. Practising alone cannot tell you everything about how they experience the work.
Should I practise privately before starting a podcast?
You can test your recording setup, rehearse an introduction and try asking questions aloud. Give that practice a specific purpose. Decide what you need to learn before recording the first episode so rehearsal doesn’t become an indefinite waiting room.
How do I get comfortable hearing my own filler words?
Review a short section for a defined purpose rather than treating the entire recording as a judgment of your ability. Notice where the wording affects clarity and choose one thing to practise. You can work on communication without making the elimination of every filler word a condition for publishing.
Can AI help me reduce filler words?
An appropriate tool may help you locate examples and design practice, but check the recording against any transcription or analysis. Some transcripts clean up hesitations. Counted words also won’t tell you everything about whether a conversation is engaging or clear. My own tool for this is still an idea I plan to build.
What if my first app gets negative feedback?
Ask what happened and which task the person was trying to complete. Fix important failures, look for recurring obstacles and distinguish those from preferences. You can use feedback to improve a release without interpreting every complaint as a verdict on whether you should build apps.
How can I use AI without overplanning?
Give each conversation a concrete deliverable and an action that follows it. Ask for one practice question, one small experiment or one piece of feedback to apply. Finish that action before requesting another round of planning unless new information makes it necessary.
How do I know when I’m ready enough?
You understand the promise you’re making, can meet the basic requirements and have a way to learn from what happens. Ready enough will look different for a podcast, a personal tool and a product handling customer data. Judge it against the task, not against whether all nervousness has disappeared.
I’m still becoming a better podcast host by hosting the podcast.
I’m noticing things I want to change because I’m doing the work. Sometimes the noticing feels uncomfortable. Sometimes it gives me an idea for something I want to build.
You can hear that process in App Minded.
Whatever your version of this is, choose the next useful attempt. Give it a finish line. Let it teach you something you can use next time.
Sources and further reading
- Shawn Wang: Learn in Public, an essay on sharing work and lessons while developing a skill.
- Public founder discussion: Building was easy. Selling feels scary., an example of hesitation at the point of visibility, not population-level research.