HN "Who Wants to Be Hired" Is a Honeypot: Source It Anyway

On October 5, 2026, a developer who posts on HN as "mrfoo" (flrnd on dev.to) dropped a minimal entry into the October "Who Wants to Be Hired?" thread with a DuckDuckGo masked address. Three emails arrived. None was a job. If your outbound plan this month involves scraping that thread, you are now the fourth email.

The post landed inside a growing HN meta-conversation about recruiter-platform spam harvesting the comment list to bootstrap AI sourcing products. The thread still has signal. The inbox around it does not.

What actually hit flrnd's inbox on October 5

Zero of the three emails flrnd received were real job opportunities. They were, in order:

1. An offshore interview-fraud pitch. An EU-based "profile sharer" offering to sit interviews using his identity; if hired, salary would route to the EU bank account, with a cut kept and the rest wired back.
2. A growth-hack community invite. A disposable address pitching a "FREE International Community for Hyper Talented Tech Professionals, Entrepreneurs and Freelancers." Reply YES to join.
3. A reverse pitch. A stranger sending his own CV, asking flrnd about remote sales roles, confusing the direction of the thread entirely.

flrnd's own summary: "Three emails, yuhu! I must be the luckiest person on the planet." The thread has a published rubric that says "Agencies, recruiters, job boards, and so on, are off topic here." That rule applies to posts inside the thread. It does not reach the inbox, which is where the AI products moved the contact.

The spam/real ratio flipped from 3:1 to 3:0 in seven years

The baseline matters. In a prior "Who got hired from HN?" retrospective, one commenter reported landing their 2019 job via WWtbH and getting one real company plus two or three spam emails alongside it. That is roughly a 3:1 spam to real ratio. flrnd's October 2026 sample is 3:0.

A single sample is a single sample. But every other datapoint in the thread's own comment history points the same direction:

• gabrieledarrigo, August 2026 thread: "Please don't spam me with your recruitment platform; contact me only if you are directly searching for a senior software engineer position." The disclaimer is now part of the post.
• Akronymus, February 2026 thread: "Use the email on my profile, the email I posted here is now getting tons of spam." He abandoned the address.
• May 2026 thread: a public call-out of nspire.ai and its named head of sales, Chris Kinnard, for spamming posters.

3:0

Why candidates are hardening their inboxes

Masked-email adoption is now the dominant signal of a serious candidate on HN. Posters using duck.com, fastmail masks, proton.me aliases, and passinbox.com show up across the October 2026, February 2026, and June 2025 threads. These are the people most committed to filtering. They are also the hardest to reach with template outreach. The highest-intent candidates have the highest-friction inboxes, which inverts the usual sourcing assumption that engagement correlates with reachability.

The pool math: 571,838 engineers, 23,215 recruiters, one HN thread

The HN thread is a Schelling point for spam because it is one of the only places on the open web that pre-labels a person as "open to work" with a contactable address attached. The arithmetic is brutal.

Roughly one recruiter per 25 engineers in the US, and automated outbound now makes each recruiter effectively louder than they were five years ago. The 120 posters on October's thread are less than 0.04% of the US software engineer pool. When a slice that thin wears an "open to work" sign, the sign is the problem.

The spam is not from recruiters. It is from recruiting products.

Read flrnd's three emails again. None came from a company trying to hire a specific engineer. The identity-fraud ring, the community-growth operator, and the reverse-pitching jobseeker are three distinct non-recruiter categories wearing recruiter-shaped packaging. A sourcer competing with "AI recruiter spam" is actually competing with these. The bar is lower than it looks, if you clear it on purpose.

The failure mode of AI recruiter outreach on this thread is predictable:

• Scrape 120 posts, extract 120 addresses, enrich against LinkedIn.
• Fan out 120 identical "Hi {first_name}, your {primary_language} background is a strong fit for our {company} opening" messages.
• Hit a wall of masked inboxes and pre-written "do not spam me" disclaimers.
• Three replies, zero of them positive.

The volume of touches went up. The quality of the match did not. That is the whole complaint arc in one sentence.

What candidate experience cold email looks like when you are not the fourth email

The outreach that actually gets a reply on an HN-sourced candidate reads like a hiring manager wrote it to one specific person, because that is what it is. The format is boring and it works:

1. Quote the post. Paste the two or three lines of their WWtbH entry that match the role. Not a paraphrase. The actual text.
2. Name the company and the team, in the subject line. "Backend role at {company}, your Elixir note" beats "Opportunity for Senior Engineer" by a wide margin on open rate.
3. Explain why the match is specific. One sentence. "You said you want to leave adtech for infra; this is infra and the team has three ex-adtech engineers who did the same move."
4. Offer a single, low-cost next step. A 20 minute call with the actual hiring manager, not a 45 minute "intro chat" with a sourcer.
5. Do not attach a job description. Link it. Keep the email short enough to read on a phone.
6. Send once. Follow up once. Then stop. The posters who get dozens of emails in 48 hours notice everyone who sends three.

The subtext of flrnd's post is that the thread is still a decent sourcing surface for anyone willing to behave like a human. The signal has not degraded. The channel around it has.

Repeat posting plus niche skills still converts

From the July 2026 thread, an exchange worth memorising. "Do people actually find jobs this way?" Answer: "I've also been contacted by a few people across the years. I post about the same text every 2-3 'who wants to be hired'. It's a bit niche though." The posters who convert are repeat posters with specific stacks. The posters who get spammed into deleting their address are the generic full-stack entries. If you are sourcing on hacker news, filter for the repeat niche posters first. They are the people the thread was designed for.

Why the thread's own moderation rule stopped working

Hacker News moderation applies to what you post, not to what you do with what someone else posted. The thread's "agencies, recruiters, job boards, and so on, are off topic" rule is enforced by flagging posts inside the thread. AI products do not post inside the thread. They scrape it, enrich it, and move the recruiter contact to the candidate's inbox, where the enforcement surface is zero.

The enforcement surface is zero because the recruiter never had to show up in the thread to begin with.

This is also why "just add a captcha" and "just hide the emails" proposals from the HN comments never land. The thread is a public artefact; the moment it is published, scraping is cheap. The only durable candidate defence is masked email plus a filter on the mask, which is what the serious posters have already done.

How to source the thread without being the fourth email

Treat the October WWtbH thread as a seed list, not a contact list. Pull out the posters whose stack, geography, and seniority actually match a live role, then reach them through a channel they chose to maintain, not through the address they are about to throw away.

A workable loop:

• Extract the 120 posters' handles and the body of each post. The handle is the durable identity. The body contains the stack, location, and intent.
• Join to GitHub, LinkedIn, and personal site via the handle. Most HN power-users use the same handle across all three. flrnd, mrfoo, and flrn.dev are the same person.
• Score for match. Role fit, location fit, "repeat poster" flag, stack rarity.
• Reach via the channel they maintain. A GitHub profile that lists an email is a maintained channel. A duck.com address in an HN post is a disposable one.
• Send a one-to-one message that quotes their post.

This is the exact shape of the workflow Refolk was built for: describe the person in plain English and the shortlist comes back with the handle, the matched text, and the best contact that is not the masked one. The output is a 15-person list you can message by hand in an afternoon, not a 120-person list you fire a template at.

The same dynamic hits every other "open to work" signal

The HN thread is the most legible version of a problem that applies everywhere a candidate raises their hand in public: #open_to_work LinkedIn frames, GitHub bios with "available for hire", conference attendee lists, Slack community intros, the "I'm looking" sub-genre of X posts. Every one is now a Schelling point for automated outbound, and every one is getting the same treatment the October WWtbH thread got.

The counter is the same in all cases: narrow the slice before you send. If the list is long enough to template, the list is wrong.

FAQ

Is posting in the HN "Who Wants to Be Hired" thread still worth it as a candidate?

Yes, if you post with a masked email, a specific stack, and a short, filterable description, and you repeat across threads. The posters who convert are the ones with niche skills who show up every two or three months. The generic full-stack posters get spammed into switching addresses. flrnd's experience is a warning about outreach quality, not a reason for candidates to abandon the thread.

What is the fastest way for a sourcer to tell a serious HN poster from a casual one?

Look at the email domain and the post history. A duck.com, fastmail mask, proton.me alias, or passinbox.com address signals someone who takes inbox hygiene seriously, which almost always correlates with a real intent to move. A post history that includes two or three prior WWtbH entries with similar text signals a repeat niche poster, which is the subpopulation that actually gets hired via the thread.

Does AI recruiter outreach work on this channel at all?

Not as currently deployed. The three emails flrnd got on October 5 are a snapshot of what the automated layer is producing: identity fraud, community growth hacks, and reverse pitches. Candidates have adapted faster than the tools. AI is useful here for finding and ranking the right 15 posters out of 120, not for sending the message. The message still has to be written by a person who read the post.

How do I reach a candidate who used a masked email in their HN post?

Pivot off the handle. HN handles overlap heavily with GitHub, LinkedIn, dev.to, and personal domains. The posters who care enough to mask their inbox usually maintain at least one of those as a durable contact surface. Message them there, quote the HN post so they know how you found them, and keep it to one send plus one follow-up. That is the shape of candidate experience cold email that still works on a channel this picked-over.