HN's May Candidate Thread Is Live. 321 Recruiters Are Mining the Wrong One.
The May 2026 "Who Wants To Be Hired" HN thread is a self-declared sourcing goldmine. Here is how to mine it before the spam wave arrives.
The May 2026 "Ask HN: Who wants to be hired?" thread (item 47975570) went live about five days ago. Every aggregator, Chrome extension, and side-project recruiter tool is currently parsing the thread next door, item 47975571, where 321 companies are competing for attention. The candidate thread, where engineers publish their stack, location, remote preference, and email in a fixed format, sits largely untouched.
That asymmetry is the trade of the week.
The two threads are not symmetric
HN's monthly hiring ritual is actually three threads: "Who is hiring?", "Who wants to be hired?", and "Freelancer? Seeking freelancer?". The first one has an entire ecosystem built around it: hnhiring.com, whoishiringjobs.com, nthesis.ai's parser, hnjobs.emilburzo.com, dheerajck.github.io/hnwhoishiring, and a Chrome extension called HN Hiring Pro. The top comment of the May 2026 hiring thread literally links to several of these tools.
The candidate thread has, prominently, one: hirehackernews.com, run by AE Studio, which pitches itself with the line "The best developers waste time on HN. Find them quick and efficiently as they wasting time." Even that tool carries an opt-out notice.
One side of the market has ten tools. The other has one. The candidates publish more structured data than the companies do.
The candidate thread's format is the gift
Every top-level reply in "Who wants to be hired?" follows a stable schema that has not changed in years. January 2026 (item 46459274) used it. May 2026 uses it:
- Location
- Remote
- Willing to relocate
- Technologies
- Résumé/CV
That is six fields of self-declared structured data, with the email address handed to you. A parser written once works next month and the month after. Compare this with LinkedIn, where you infer "open to work" from a frame around a profile picture and infer stack from a self-written headline that was last updated in 2023.
Why "no recruiters allowed" is a feature
The thread states: "Agencies, recruiters, job boards, and so on, are off topic here. Please also avoid recruiter speak, e.g., 'We are looking for passionate, talented developers who want to join our amazing team of disruptors.'" It then asks readers to "only email these addresses to discuss work opportunities."
Most recruiters read that and move on. Good. That gate is doing free work for the few who stay. Response rates collapse when 40 people in the same week paste the same templated InMail. They hold up when the candidate has been emailed three times, by three humans, each of whom quoted the candidate's actual post.
The rule is not "no one who is hiring." It is "no agencies, no job boards, no recruiter-speak." A founder writing a five-line note about why their Rust + embedded role fits the candidate's exact stated stack is not what the rule was written to block.
The "no recruiters" rule deters lazy recruiters, which is exactly why response rates for well-targeted outreach stay high.
The numbers behind the asymmetry
Cold email is in long-term decline. The Instantly benchmark puts average response rates at 8.5% in 2019, 5% in 2025, and 3.43% in 2026. Inbox saturation, smarter filters, and low-effort AI-generated outreach are all eating it.
Recruitment cold email holds up better, landing between 5% and 8% because "the pain is immediate and obvious." One tech recruiter reported 12% response on cold email vs 4% on LinkedIn InMails, with cost-per-hire dropping 60% as 70% of pipeline shifted to email.
Now layer the HN candidate thread on top. These are not passive prospects you have to convince. They typed their own email address into a public thread under the words "Who wants to be hired?". Intent is not inferred. It is published.
Industry sources disagree about InMail. One claims InMail hits 18-25% in 2026 versus cold email's 2-3%. Others put InMail at 10-25% with personalized outreach pushing 30-50%. Pick whichever you like. The point is that any of those numbers collapses when the same candidate has been hit thirty times in seventy-two hours. The HN thread has a roughly 72-hour window where the top posters are still inbox-fresh.
What you actually get per post
A representative reply in this thread tells you:
- The candidate's city and country
- Whether they want remote, hybrid, or onsite
- Whether they will relocate, and where to
- A specific stack list ("Rust, Postgres, embedded Linux, some Zig")
- A link to a résumé or personal site
- A working email address
- Often: salary band, visa status, what kind of company they want to join, and what they are explicitly not interested in
Compare with the average LinkedIn discovery flow. You search a title, filter by "Open to work" (a tiny slice; in one query against professional-network data, only around eleven senior SWE profiles across the US/UK/Germany/Canada matched "open to work" + remote), send an InMail, get a reply two days later asking for the JD, send the JD, lose them to a counter-offer in week three.
In the HN thread you skip three of those steps because the candidate already wrote the answers down.
Mirror the company thread onto the candidate thread
Here is the move most recruiters miss. The "Who is hiring?" thread is a free market-demand survey. Read it before you write your outreach.
In May 2026 the company thread is loud about specific stacks:
- IBM Research is hiring "early career scientists and engineers interested in LLMs combined with programming languages, formal methods, or compilers, with Rust experience as a plus."
- Ashby (YC '19, scaled engineering 30 to 60, planning to double) is hiring across the stack.
- Kraken Tech's AI Foundations team is hiring across London, Paris, Berlin, Tokyo, NYC, Tel Aviv, and remote.
- Prior Labs (Berlin) is hiring on tabular foundation models.
- Dimensional (robotics, $200M+ raised, team out of MIT, CMU, DJI, and Figure) is hiring.
- Burnin (ex-Meta FAIR, Pinecone, AWS) is hiring.
- OpenVPN, Qualcomm, Mobian, CodeWeavers are all in the thread.
Now flip to the candidate thread and grep for "Rust + LLM," "tabular ML," "robotics + perception," "OpenVPN" or "VPN," "Wine internals." You will find named-match candidates whose post reads like it was written for one of the JDs above. That is the trade. The two threads were posted by the same moderator, two minutes apart, in the same hour. Nobody pairs them.
This is one of the workflows we built Refolk for. You describe the person in plain English ("engineers who posted in this month's HN candidate thread with LLM and formal methods experience, ideally based in Europe"), and you get a ranked shortlist with the source post, the stated stack, and the published email. The "find" step compresses to a sentence.
A note on Canada, the EU, and the US-only gate
Inside the May 2026 hiring thread, a recurring candidate complaint: "As an engineer in Canada, I'm quite frustrated by US-only when it's a remote job. It's so easy to employ Canadians either as a contractor or through a proxy like Rippling."
Recruiters who don't gate on US-only get an unfair pick from the candidate thread. EU-based, Canada-based, and LATAM-based engineers are over-represented in "Who wants to be hired?" relative to the companies that are willing to hire them. If your client or your own company is set up for global contractors, the conversion math on HN is meaningfully better than on LinkedIn, where the same candidates do not bother turning on "Open to work" because they assume the inbound will be US-gated.
The 72-hour playbook
Treat this as a perishable opportunity, not a database.
Hour 0 to 24: parse, dedupe, segment
Pull every top-level reply. Strip the field labels. Build columns for location, remote preference, stack tokens, seniority cues, and email. Segment by your two or three live reqs. If you are sourcing across multiple roles, this is where Refolk earns its keep: plain-English filters across the HN thread plus the rest of GitHub, LinkedIn, and the open web in one query, instead of three separate scripts.
Hour 24 to 72: write a real first message
Quote one specific line from the candidate's HN post. Reference the actual stack overlap with your role. State the location/timezone/visa situation up front. Do not paste a JD. Do not say "I'd love to learn more about your background." They already wrote their background down for you.
A good first message is four sentences:
- The line of their post you noticed.
- The role and why it fits the specific thing they wrote.
- Comp band and location/remote situation.
- A one-sentence ask.
Day 4 to 14: the follow-up window
By day four the top posters are getting flooded. Your follow-up should reference your first email, add one piece of information they did not have (a teammate they would work with, a customer they would build for, a piece of the codebase), and close with the same one-sentence ask. Drop the candidate from the list at the third no-reply.
What the asymmetry means longer term
The candidate thread will get more crowded. AE Studio's tool already exists. Someone will build a slicker one. A YC batch will spit out three competing parsers next year. The window where this is a quiet sourcing surface is measured in months, not years.
The deeper point survives the window closing. Self-declared candidate signals beat inferred ones. A thread, a Discord, a Slack, a Show HN comment, a conference attendee list, or a GitHub issue thread where engineers describe their own situation in their own words will outperform any LinkedIn title search on first-reply rate, every time. Refolk's whole premise is that the find step should be a sentence, because the qualify step is where recruiters actually add value, and the qualify step is what self-declared data shortcuts.
Mine the May 2026 candidate thread this week. Then build the habit of checking the first of every month at 11am Pacific, when the next one drops.
FAQ
Is it actually allowed to email candidates from the HN "Who wants to be hired" thread?
Yes, if you are the hiring company. The thread explicitly invites readers to "only email these addresses to discuss work opportunities." What it bans is agencies, job boards, and recruiter-speak templates. A founder, hiring manager, or in-house recruiter writing a specific, personalized note about a real role is exactly the use case the thread is designed for. If you are an agency, you need a different play (work it through the hiring company, not the candidate).
How is this different from just using hirehackernews.com?
hirehackernews.com parses the candidate thread, which is useful. The asymmetry this article is about is workflow, not just data. You want to cross-reference candidates against your specific open reqs, against the parallel "Who is hiring?" thread, and against everything you already know about each person from GitHub and the open web. Refolk handles that cross-reference in one query; a single-source parser leaves you to do the joining by hand.
What is the realistic response rate on outreach to HN candidate-thread posters?
Higher than generic cold email, especially in the first 72 hours, because intent is published rather than inferred. The benchmarks worth anchoring on are the 5 to 8% range for recruitment cold email overall and the 12% reported by one tech recruiter on cold email vs 4% on InMail. Expect the top quartile of HN posters (specific stack, clear remote preference, named target companies) to outperform those numbers materially if your first message quotes their post. Expect the bottom quartile (vague, "open to anything") to underperform.
When does the next thread drop and how should I prepare?
The first weekday of each month, around 11am Pacific. Whoever posts "whoishiring" runs both threads back to back, two minutes apart. Prepare by keeping a live list of your open reqs translated into plain-English candidate descriptions ("senior Rust engineer, EU remote, embedded background, has shipped a compiler or runtime"), so that on day one you can paste those descriptions into a sourcing tool against the new thread and have your shortlist before lunch.