Distribution
Hacker News Cannot Be Automated, on Purpose
Hacker News is the one major platform where automation is structurally impossible, not merely discouraged. There is no write endpoint in its official API, its robots file disallows every write path to automated agents, and its guidelines separately state that generated text should not be posted at all. Each barrier alone would stop most tools.
Three separate walls, not one
Most platforms that want to slow down automation pick one lever and pull it. A rate limit. An approval process. A terms of service clause nobody reads until they get banned.
Hacker News did something different. It built three independent barriers, each one sufficient on its own, stacked on top of each other.
The first wall: there is nothing to post through
Start with the official API, published on GitHub and maintained by Hacker News itself. Read the endpoint list and a pattern jumps out immediately: every single one is a GET request. Item lookup, user lookup, top stories, new stories, the maximum item ID, an update stream. There is no POST /submit, no POST /comment, nothing that writes.
The README describes the API as, in its own words, essentially a dump of the site's in-memory data structures. That framing matters. This was not built as a product for third parties to publish through. It was built so people could read what is already there. Writing was never in scope.
The second wall: the door is marked closed
Suppose you decided to skip the API entirely and just script a browser to fill in the submission form on the actual website, the way a person would.
Hacker News's robots.txt addresses that too. It disallows every write-shaped path by name: /submitlink?, /vote?, /reply?, /flag?, /fave?, and a handful of others, each one a page you would need to touch to post, vote, or comment. It adds a thirty second crawl delay on top, which by itself would make any scaled activity impractical even where it is technically permitted.
This is worth sitting with for a second. A robots.txt file is usually a polite request, not an enforcement mechanism. Hacker News wrote a fairly specific list of exactly the actions a marketing bot would want to take, and asked automated agents not to take them. That specificity is a decision, not an oversight.
The third wall: the content itself is disqualified
Here is the part that makes Hacker News different from every other platform in this category. Even if you found a way around the first two barriers, the guidelines close the last door directly, addressing the words themselves rather than the mechanism that produced them.
The line reads, without qualification: "Don't post generated text or AI-edited text. HN is for conversation between humans."
Read that again. It does not say generated text needs a disclosure. It does not say it needs approval. It says do not post it. A founder who somehow got a script through the API gap and around the robots restriction would still be breaking this rule the moment the words themselves came from a model rather than a person.
What this rules out, precisely
Put the three walls together and here is what falls out of scope entirely: scheduled submissions, automated replies to comments, any workflow where a model drafts a comment and a script posts it, and technically even a person copying a model's draft into the box without substantially rewriting it in their own words.
What survives is narrower and slower. A person who reads Hacker News regularly, who has commented usefully before today, who writes the submission and every reply themselves, and who is prepared to sit with a thread for the eight or ten hours it stays active. That is not a workaround for the restrictions above. It is the only shape of participation the site was built to allow.
Why this is the clearest version of a pattern
Hacker News happens to be the strictest platform on this specific question, but the shape of the problem shows up everywhere distribution actually works. Reddit gates its API behind approval and a separate commercial agreement, and lets any subreddit's moderators remove a technically compliant post anyway. Even LinkedIn, which does allow sanctioned API posting, distinguishes carefully between an authorized integration and an unauthorized bot, and the distinction is doing real work.
Hacker News just removed the ambiguity. There is no gray area to argue about, no policy to interpret generously. It is either a person, writing as themselves, or it does not happen. That is an unusually clean example of the thing every serious distribution channel eventually enforces in some form: the account has to be a real participant, because the entire value of the platform depends on that being true.