New business ideas: what builders launched in 2026, and who noticed

New business ideas are not scarce. Attention and buyers are. Our Discovery runs read 17 Hacker News threads posted in 2026, and 11 of them were product launches. None of the 9 ordinary Show HN launches reached 5 points. Below is every problem in those threads, what was launched against it and how much attention it got — and why the useful question about a new idea is not whether it is new, but who will answer.

Where these threads come from

Our runs reach threads like these while they search a market for problems. That makes this a sample of what the runs happened to read, not a survey of Hacker News. Every thread is dated by Hacker News's own public API — they run from 2026-01-03 to 2026-08-10 — and the points and comment counts are a snapshot taken on 2026-09-27.

Two things to know before reading the table. The runs saw each of these through a search result, not by opening the thread, so the problem column is the run's reading of a snippet. And many of the threads are launches: a builder describing the problem their own product solves, which is a different kind of evidence from a stranger complaining. Each row says which kind it is.

Most launches were not noticed

Here is how much attention each launch got:

A quiet launch does not prove nobody has the problem. It proves that one post, on one day, did not reach anyone who cared enough to click. That is still the most common outcome in this sample, and it is the part a list of new ideas never shows you: the builders here had the idea, built the product and published it, and then met silence.

The two loud launches say something too, and it is not only about their ideas. A Launch HN post is a slot only Y Combinator companies get, with placement an ordinary Show HN does not. Their bars measure distribution as much as demand.

The problems, and what was launched against them

Swipe left to see what appeared on Hacker News and how much attention it got.

Where the better evidence is

The launch rows tell you what someone chose to build. The discussion rows are more interesting, because nobody in them was selling anything. A complaint that meeting notes are recorded and never read again, made in passing under a story about a meeting-notes company's security lapse, is closer to real demand than a product page describing the same problem. It is still one person, and it still needs a buyer who will pay — but it is the kind of signal worth chasing.

Some of these problems have been tried many times. Heavy time-tracking and invoicing tools for solo freelancers appear here as a quiet launch, and a version of the same problem sits on our shortlist of business ideas that survived a filter. A crowded problem is not disqualifying. It means your version needs a narrower buyer, a reason to switch and a channel the last builder did not have.

How to turn one of these into your idea

1. Pick a problem whose buyer you can reach. The second column is the run's guess; if you cannot name ten of those people, the problem is not yours yet.
2. Open the thread. The run only saw a snippet. The comments tell you who cares, what they use today and what they already said no to.
3. Write the narrow version. One buyer, one outcome, one price — narrower than whatever was launched against it.
4. Ask for money before you build. A launch post collects attention; a pre-order or paid pilot answers the price question. What each signal can prove walks through the difference.
5. Decide what silence means before you post. Most launches here were met with it. Yours should have a kill line written down in advance.

If none of these problems fits the time, money and skills you actually have, a list will not fix that. A founder read will. If you already have an idea of your own, check that one idea instead.

The short version

• New business ideas are cheap. In this sample: None of the 9 ordinary Show HN launches reached 5 points.
• The two loud launches were Y Combinator posts with placement ordinary builders do not get.
• The best evidence here is a problem raised in passing in a discussion, where nobody was selling.
• A crowded problem needs a narrower buyer, not a better pitch.
• Ask for payment before you build, and write down in advance what silence means.