Hacker News “Who is hiring?”: how to actually search hundreds of comments

What makes the "Who is hiring?" thread hard to search?

It is a top-level Hacker News post where anyone replies with an opening, and anyone can reply to those. No login wall, no form between you and the person who wrote it. You read the real words of someone trying to hire, which is rarer than it should be.

That is also the problem. Hacker News's own search API reported 393 comments on the October 2026 thread when it was read on 19 October 2026. Replies are included, and 29 were hidden from logged-out readers. Of the rest, 266 were top-level posts, all on one page, in the order Hacker News ranks them: its FAQ says comments are ranked the same way as stories, from points and age plus other signals. On 19 October the ten newest top-level posts sat anywhere from 4th to 257th place. A new post can be anywhere.

The same search API reports these totals for earlier October threads:

Every count includes replies, and posting habits have changed over the years. The table shows how much reading one thread takes, and says little about hiring.

There is no filter for "remote", none for "backend", none for "posted in the last hour". Replies carry questions, "is this still open?" messages from weeks later, and the odd repost. And there is no applicant count anywhere, so there is no reliable way to infer your place in the pool or a review advantage from when you respond. I use the post date as a freshness signal, not as a promise of recruiter preference. The same limit applies to public applicant counters on LinkedIn.

Which Ctrl-F searches work?

Before anything fancier, find-in-page (Ctrl-F or Cmd-F) goes further than it should, if you search for the thread's own conventions instead of your own words.

1. REMOTEThe most useful search in the thread. Its rules ask posters to write REMOTE in caps, with the country in brackets if it is restricted. In the October 2026 thread, 76 of the 266 top-level posts did, and 150 mentioned remote in some form. Chrome's find bar ignores case, so a search for REMOTE lands on all 150. Firefox has a Match Case option that keeps you to the 76 tagged posts, and misses the ones written as "Remote".
2. ONSITEThe inverse, for ruling postings out fast if remote is non-negotiable.
3. A locationA city or country, if you are anchored somewhere and want to skip the flood of US-only listings.
4. $Posts that state a salary or range. It is a much smaller set, about one top-level post in four in October 2026, and worth reading first. It also catches funding rounds, so skim past those.
5. Top-level comments only, on the first passReplies under a posting are nearly always questions or clarifications, rarely a second listing. Reading every chain is how an hour disappears.

The bracket after REMOTE is where the catch usually sits. In the October 2026 thread, 60 of the 76 REMOTE posts added one, usually naming a country or region, and 28 of those brackets began with the US. Oink's own store of postings shows a similar lean. Of the 83,780 remote postings in it on 18 October 2026, 87.1% said who may apply, and among those naming exactly one country or region, 65.6% named the US. Those figures cover only Oink's store, a slice of the whole market. So search for your country right after REMOTE. Then read the bracket before the rest.

Are there tools for this?

Enough people hit the same wall that a few tools exist. Hacker News's own Algolia-powered search can be scoped to a single story, which gives the thread a query box. Community "hnhiring"-style trackers pull each month's thread into a searchable list, with tags for remote, visa sponsorship or tech stack taken from the free text. The thread points to some of them itself: the October 2026 post lists five search tools under "Searchers" in its opening text.

Can you pull the thread programmatically?

Yes. The thread is a Hacker News story with comments, public and structured enough to fetch by script. The official Hacker News API stores stories and comments as items with IDs, timestamps, text, parent links and child IDs, and your code has to walk the comment tree itself. Once the comments are structured, each Ctrl-F pattern above can be a real query. Deleted or edited items and inconsistent free text still need handling.

Who does the reading?

Scraping solves finding. It doesn't solve reading: someone still has to open each match and decide whether it is worth your time, and a keyword filter can't make that call.

Oink is built to cut down step two. Hacker News can take part next to broad boards, specialist sources, employer pages and open-web discovery, when those paths are on and healthy. Viable candidates are ranked from 0 to 5 with the reasons they fit and the concerns. Which sources take part, and how much detail each record has, can vary. The job-source guide avoids a fixed count and lists the ways a source can come back empty.