A new analysis of Hacker News posts and comments argues that software availability does not necessarily satisfy user demand: 81 of the 100 most frequently requested product ideas in the project’s ranking have already appeared on Show HN, yet people continue to ask for them.

The project, called Still Unbuilt, says it collected 63,300 posts and comments dating to 2008 that used phrases such as “I wish there was,” “is there a tool” or “I’d pay for.” An artificial-intelligence classifier then identified 12,441 entries as software requests. The site groups and ranks the resulting ideas by the number of separate discussions, the years in which they appeared and recent demand. Links are provided to underlying Hacker News discussions.

Privacy, local control and cleaner versions of existing services feature prominently. According to the project, 46 of the top 100 requests predominantly seek local or private software, while only eight chiefly ask for artificial intelligence to perform the task. Four of the top 10 include an expressed willingness to pay for products such as a stripped-down Windows edition, a browser, an email service and an advertising-free search engine.

The ranking’s recurring requests include tools to filter unwanted material from feeds, a Firefox-based browser without advertising or telemetry, private email with conventional standards support, and a clean paid search engine without tracking. Other repeated ideas include reliable cross-platform mail clients, encrypted backup software, simple RSS readers and private personal knowledge stores.

Many of these categories already contain numerous products. The analysis counted 1,423 Show HN launches across the 81 top-ranked ideas it classified as built. Its examples include dozens of launches for content-management systems, personal knowledge tools and article summarizers. Repeated requests may therefore signal dissatisfaction with price, design, trust or maintenance rather than the complete absence of software.

The project also reports a shift toward requests for improved versions of familiar products. It says 27% of requests since 2024 seek a better version of existing software, compared with 14% in the 2008-to-2011 period. Sixty-four of the top 100 ideas were first requested in 2012 or earlier, suggesting that some preferences have persisted through multiple product cycles.

The results should be read as an exploratory analysis of one technology community, not a representative survey of software buyers. The collection method depends on selected phrases, and the classification and grouping were performed by an AI system. Still, the linked discussions make the database a useful map of recurring complaints. Its clearest finding is that launching a product and resolving the problem are different things: users continue asking when existing options fail their expectations for control, simplicity, privacy or value.