從自己的一週尋找公開建構內容靈感，而不是依賴 27 個點子清單

the best build in public ideas are not on a list of twenty-seven. they are in your own week: the thing you shipped, the decision you went back and forth on, the customer who said something you did not expect, the question people keep typing that nobody answers well. lists of post ideas are fine as categories. they are the same list for everyone, which is the problem, and they run out the moment you have used each one once.

the short answer: pick ideas from four places only you have. what happened since you last posted. what people search for that you could answer. what an ai assistant says when your buyer asks it about your market. and what your own marketing has not said yet. then do them in order, three at a time, and judge it by totals that grow, not by your best post.

below: the usual categories and why they stall, the four sources with examples of each, twenty ideas you can take straight from them, how to choose, and what underway's today page does with all of this.

why lists of build in public ideas run out

page one for this search is mostly lists sorted into the same five buckets: progress updates, decisions, failures and lessons, metrics, and stories about the people you serve. the buckets are right. they describe what good build in public posts are made of. what they cannot do is fill a bucket for you on a tuesday when you are tired and nothing feels worth saying.

there are two other problems. a generic idea makes a generic post: "here is my weekly shipping recap" reads the same from everyone. and a list does not say which idea matters today. twenty-seven options with no order is a list of things you will not do. what helps is fewer ideas, each specific to you, with the reason it is worth doing now.

four places to find build in public ideas that are yours

what happened since your last post. a payment landing, a call that went sideways, a commit you were stuck on for three hours, a document you rewrote. the tools you already use usually know before you remember.
what people search for that you could answer. a phrase people actually type into youtube or google, where page one is thin, old or off the point, is an idea with an audience already attached.
what an ai assistant says about your market. ask it the question your buyer would ask. if it names three other tools and not you, the idea is the video or page that answers that question plainly.
what you have not said yet. write down who it is for, what they do instead today and why you are different. the gaps, the lines you cannot fill in, are the next things to say out loud.

none of these is a template. each one is specific to you, which is why it is worth doing, and each one refills itself: there is always a new day, a new search, a new answer to check.

the kind of moment that beats any list: something on the board is not going the way you hoped. an ai illustration, not a customer.

twenty build in public ideas from those four places

each of these comes from one of the four sources above. none needs a plan; each needs one honest answer.

the thing you shipped this week, and the part of it nobody will notice
a bug that took far longer than it should have, and what it actually was
a decision you reversed, and what changed your mind
a price you set or changed, and how it felt to press save
a meeting that changed the plan, said without naming anyone who did not agree to be named
a customer message that surprised you, paraphrased
the feature you decided not to build
a quiet week, said honestly
the question a customer asked that your website does not answer
a search people type about the problem you solve, answered to camera
a how-to for a job your product does, shown done with it
what people use instead of you today, and where it falls short for them
an ai assistant's answer to your buyer's question, and what you would add
who it is really for, now that you have talked to more of them
the hardest part right now
the goal you set at the start, and where you honestly are against it
a number that moved, as a total, without making it a contest with last week
something you tried in marketing that did not work
a tool you connected or dropped, and why
what you would tell yourself on the day you started

how to choose: an order, not a list

a short rule works. first, anything that happened since your last post, because news goes stale and a serial is about what happened next. then anything with a search behind it, because it keeps finding people after the week is over. then the gaps in what you have said, because they make every later post easier to understand.

and keep it to three. three things in order, each with a reason, and the rest out of sight until you want them. dismissing an idea should cost nothing: no reminder, no guilt, no second ask.

	a list of post ideas	ideas from your own week
where they come from	someone else's categories	what happened, what people search, what ai answers say, what you have not said
how many	twenty-seven, all at once	three, in order, each with why
how long they last	once each	they refill every week
what you post	a format	something only you could say

how to tell build in public ideas are working

not by your best post. ranking your own posts makes you chase your best day, and that is how a diary turns into hook-chasing. look at totals instead: how many episodes or posts you have made, how many people follow you across the places you post, how many searches you are on page one for, how often an ai assistant names you when asked about your market. totals go up when you keep going, which is the only thing that was ever going to work.

consistency is the other half. post as often as you said you would, and no more. a commitment you can keep on a bad week beats one you break on a good one.

what underway's today page does with this

underway is a video diary for founders: once a day, whenever suits you, it asks you about your work and cuts the answers into a numbered episode. today still opens on the record button, first and largest, because that is the habit. under it are two things built from the four sources above.

your progress: episodes so far, people following you, searches you're on page one for, and ai answers that name you, each with the week's change. a tile stays off until it has a number. totals only, never a ranking of episodes.
what's worth doing now: the three things most worth doing, each with why, and show all for up to ten. they come from your ai answers, videos to film, your marketing brief and what it couldn't tell (answer those right there), and tools worth connecting.
each step has a button that takes you there, and not now, which puts it away for good, on the web and the phone. a part of underway you have not opened yet comes back a week after not now, until you open it.
none of it reaches your interviews. what you are asked on camera comes from your day; the next steps live beside the diary, not inside it.

today on the phone, record button first. this shot is from 27 september 2026, before your progress and what's worth doing now were added under it.

where the ideas come from inside underway

what happened: connected tools (github, calendar, slack, stripe and others) land as lines in today's context, and the director ticks the ones that matter to your season. the interview's questions are written from those.
what to film: in the setup, one list of topics from your season and from what you have said on camera, tagged since your last episode, left open, new, or what the story needs next.
what people search: videos to film, in promote, is what your audience types into youtube, from youtube's own suggestions, never padded, with how to film each.
what ai answers say: found in ai answers, on the series plan, asks your buyers' questions of ai assistants once a month and shows whether they name you.
what you have not said: your marketing brief, on every plan, and what it couldn't tell.

the walls matter. search research and ai answers are shown to you in promote and as next steps; they never reach the interview, its questions or the cut. the diary stays a diary. the shop window sits beside it.

if you want the ideas pulled from your week for you, the apps compared: build in public app.

what are good build in public ideas?

the ones only you have: what happened since your last post, questions people search that you could answer, questions an ai assistant answers without naming you, and what your marketing does not say yet.

what should i post when nothing happened?

the quiet week, honestly, or a gap in what you have said: who it is for, what people do instead today, the hardest part right now.

how many build in public ideas should i work on at once?

three, in order, each with a reason. a long list without an order is a list of things you will not do.

where do underway's next steps come from?

from your ai answers, videos to film, your marketing brief and your connections. nothing generic.

what happens when i press not now?

that step goes for good, on the web and the phone. a part of underway you have not opened yet comes back a week later, until you open it.

do the next steps change my interviews?

no. they never reach the interview, its questions or the cut.

for the weeks that feel empty: what to post when nothing happened.

the search half of the ideas: how to find what people search on youtube.

Originally published at https://underway.life/build-in-public-ideas