Hacker News Is Not Product Feedback

The top comment on your Show HN optimizes for an audience that will never pay you.

The hook

A team I work with posted a Show HN for their developer CLI tool three months ago. Front page. Two hundred replies in the top thread debating whether it should have been a VS Code extension. By Monday, the product roadmap had changed. A Jira epic called "VS Code Extension Investigation" existed, sourced entirely from the HN thread. I asked the lead how many paying customers came from Hacker News. He did not know. I checked. It was four.

निंदी कोणी मारी · Let Some Condemn, Let Some Worship

निंदी कोणी मारी । वंदी कोणी पूजा करी ॥
मज हे ही नाही ते ही नाही । वेगळा दोहीं पासुनी ॥
The person who is separate from both praise and blame is free to act on what they actually see, not on what others say about them.

What I keep seeing

The second HN thread was pure praise: "beautiful README," "love the ASCII art," "finally someone gets developer UX." The third was a tear-down: "just a wrapper around curl," "why does this exist when httpie already does this." The designer started redesigning the onboarding flow because one commenter said first-run experience was confusing. Nobody checked whether their actual paying users, about three hundred teams on a $29/month plan, had ever requested any of these changes.

I have fallen for this myself. Early in my career I rewrote an entire authentication flow because a Reddit thread called it "clunky." Our paying users had never mentioned it. Not once. The new flow introduced two bugs and a three-day outage. The Reddit commenter never signed up.

Here is the uncomfortable part: the HN praise is just as dangerous as the criticism. The team started protecting the ASCII art in the CLI because strangers liked it, even as paying users asked for machine-parseable JSON output instead.

The mechanics

Hacker News has roughly 500,000 unique daily visitors, mostly developers and founders. The voting and commenting system selects for contrarian takes, novel framing, and strong opinions. A comment saying "this is fine, does what it says" gets zero upvotes. A comment saying "this already exists, it is called X, and X does it better" gets fifty. This is not a bug; it is the mechanic that makes HN interesting to read. But it means the signal from a front-page post is filtered through a selection function that rewards disagreement and novelty, not utility assessment.

Your paying users found your tool through a Google search, a colleague's Slack message, a Stack Overflow answer. They evaluated it against their workflow. They entered a credit card number. Their feedback, through support tickets, NPS surveys, churn interviews, is indexed to whether the tool solves their problem. HN feedback is indexed to whether the tool is interesting to discuss.

Treating them as the same signal is like tuning your recommendation engine on Twitter impressions instead of purchase conversions. The metric is real. It measures the wrong thing.

Where Tuka comes in

Tukaram does not say "ignore criticism and accept praise." He does not say "ignore both." He says मज हे ही नाही ते ही नाही: for me, neither this nor that. A third position. Separate from both responses, capable of seeing the situation without the distortion that praise and blame introduce.

This is an epistemological claim, not a stoic posture. The person who absorbs the praise starts protecting the praised elements from change, even when the product needs them to change. The person who absorbs the criticism starts justifying their existence instead of improving their product. Both responses move the builder's attention from the thing being built to the audience watching. वेगळा दोहीं पासुनी (separate from both) is the position where product decisions come from what you observe in your users' behavior, not from what strangers say about you on a forum.

What I would actually do

Post on Hacker News for distribution. Never let the comments touch your roadmap. Write a policy, visible, referenced in planning: HN threads go into a "public sentiment" log that marketing reviews for positioning language, not into the backlog. Product decisions come from three sources only: support tickets from paying users, usage analytics from PostHog or Amplitude, and direct interviews with churned customers. If an HN comment matches something paying users already say, that is confirmation. If it does not match, it is content for someone else's product. File it. Move on.

Chetan Dhandal