1,200 Amazon Engineers Signed an Open Letter. LinkedIn Sees 36.

On June 3, 2026, three Amazon engineers stood at a Seattle City Council hearing and criticized their employer's $200B AI capital spend while the company shed 30,000 corporate roles. A week later, Amazon Employees for Climate Justice (AECJ) published an open letter with more than 1,000 named Amazon signers demanding worker input on AI-driven layoffs. If you are sourcing engineers, that letter is the highest-intent candidate list you will see this quarter.

Most recruiters will read the news, nod, and go back to filtering LinkedIn for "past company: Amazon." That is the mistake this piece is about.

The letter is a sourcing artifact, not a news story

AECJ's open letter is a publicly named, role-tagged, team-tagged roster of 1,000+ current Amazon employees who have declared, in writing, that they disagree with their employer's strategy. That is a stronger signal than any "open to work" badge and it is sitting on a static URL right now.

Here is what is actually on the page:

• Named engineers with titles like "Sr. SDE, Alexa AI," "SDE, AWS Batch," "Software Dev Engineer III," and "Applied Scientist."
• Team affiliations LinkedIn almost never surfaces: EC2, S3, Amazon Robotics, Data Center Automation, AWS Agentic AI.
• A mix of levels and functions: high-ranking engineers, applied scientists, economists, senior product leaders, marketing managers, and warehouse staff.
• A parallel solidarity list of 2,400+ signers from Google, Apple, and other companies.

AECJ committed to publish job roles and orgs only after collecting at least 1,000 signatures. They hit that threshold. The org-level detail is now the point.

1,000+

Why "sourcing amazon engineers" from LinkedIn misses this

LinkedIn scraping surfaces people who have already left. It does not surface current employees who are actively unhappy. Revealed preference beats inferred preference, and the AECJ letter is the cleanest revealed preference artifact in tech right now: a name, a team, and a public act of dissent, all on one page.

The Refolk index says the "AI critic" pool is 28x smaller without the letter

In Refolk's index of professional profiles, only 36 people publicly match the "Amazon + Seattle + AI-critique keywords" cluster, versus 1,000+ named on the letter itself. That is a step-function difference, and it is the reason dissent artifacts beat keyword scraping.

Here is the picture side by side, drawn from the research desk and Refolk's index:

Two things fall out of that table. First, the letter is the source of truth, not LinkedIn. Second, most signers are still inside: 58% of the matchable Seattle-AI-critic pool is still at Amazon or AWS. You are not sourcing alumni. You are sourcing a pre-attrition cohort.

HR retaliation is a recruiting accelerant, not a deterrent

One day after Seattle passed its 9-0 data center moratorium on June 9, 2026, all three named engineers, Patrick Schloesser, Liesl Wigand, and Darius Irani, were pulled into an unplanned Zoom with Amazon Employee Relations. AECJ characterized it as a threat of termination. Every other signer now knows their name is on a list Amazon is watching.

That changes the math on cold outreach in a way that is easy to miss:

1. Acute loss of psychological safety. Signers went from "principled stand" to "under investigation" in a week.
2. Public precedent. In April 2020, Amazon fired AECJ co-founders Emily Cunningham and Maren Costa; Amazon settled after they filed an NLRB complaint in 2021. Every signer has read that story.
3. Named exposure. The three testifiers are on CNBC and Fortune by name. The rest are one search away.
4. A 2024 NLRB case already alleged Amazon retaliated against corporate employees who walked out over the RTO mandate. This is a pattern, not a one-off.

Reach out in the two weeks after a retaliation story breaks and the reply rate you get on a passive candidate looks nothing like your baseline. This is the window.

Signers went from principled stand to under investigation in a week. That is not a deterrent. That is a deadline.

The letter is a team map, which is what makes it rare

Most public rosters give you names. AECJ gives you names plus internal org, which is the piece you cannot buy or scrape. If you are building AI infrastructure, a "SDE, Data Center Automation" who signed a letter opposing data center buildout is a preposterously precise match for your first ten hires.

Think about how you would normally get that specificity:

• LinkedIn: title is usually "Software Development Engineer" with no team.
• GitHub: reveals stack but not internal org unless the person contributes to an OSS project tied to a specific service.
• Referrals: require you to already know someone on that team.
• Conference talks: bias toward staff+ and toward teams with a PR budget.

The letter cuts all of that. It gives you SDE III and Applied Scientist grade signal on teams that never publish. Feed that into a tool that understands plain-English queries and the workflow gets tight fast. This is the exact gap Refolk closes: describe the person in plain English, including team and dissent signal, and get a ranked shortlist across GitHub, LinkedIn, and the open web.

The non-obvious pools everyone will ignore

Three sub-pools inside the letter are worth more than the headline number, and every one of them will be underworked because recruiters read "climate" and stop.

Warehouse and non-engineering signers

Signers include warehouse staff, economists, marketing managers, and senior product leaders. If you are hiring for ops, supply chain, or GTM, you are staring at a pre-qualified list of people who have already publicly said they want to work somewhere that treats AI deployment with more caution. Founders hiring their first ops lead ignore this because they assume the letter is an engineer thing. It is not.

The 2,400+ solidarity signers at Google and Apple

The adjacent list is bigger than the primary. Solidarity signers from Google, Apple, and other companies added their names, and they are self-selecting on the same axis: skeptical of "warp-speed" AI deployment inside big tech. If your pitch is "we ship AI thoughtfully," that is your ICP, and recruiting from big tech into a smaller, more deliberate shop is exactly the swap they are already primed for.

The future-founders subset

The 2020 firings of Cunningham and Costa produced a founder cohort in climate tech and worker tooling. Expect a subset of the 2026 signers to leave in 2026 and 2027 and start climate-AI, data-center-efficiency, or worker-tooling companies. Getting in front of them at signature time is 12 to 18 months before they show up in "just left Amazon" filters. If you are a seed investor, this list is a deal flow artifact, not a hiring list.

How to actually work the list this week

Do not scrape the letter and blast it. That torches the signal for everyone. Work it the way you would work a walkout organizer list or a No Tech for Apartheid signatory list, which is to say: narrow, specific, and respectful of the reason they signed.

A concrete playbook:

1. Pull the AECJ open letter page and extract signers by team, not by name. You want clusters, not a spreadsheet dump.
2. Filter to the teams your product actually competes with or complements. If you build inference infra, "SDE, AWS Batch" and "SDE, Data Center Automation" matter. "Marketing Manager, Prime Video" does not.
3. Cross-reference with GitHub for stack fit. Amazon engineers who signed and who also maintain public repos are the highest-intent slice.
4. Reference the letter in your outreach, not the layoffs. "I saw your signature on the AECJ letter and the team affiliation you listed" beats "Amazon layoffs 2026 got me thinking about you" by an order of magnitude.
5. Skip the three testifiers. Schloesser, Wigand, and Irani are getting inbounded by everyone. The 997 quieter signers are the actual opportunity.

This is the workflow Refolk was built for. You describe the person ("current Amazon Applied Scientist on Alexa AI who signed the AECJ open letter, based in Greater Seattle") and get named candidates with the team detail already resolved.

Dissent artifacts are a category, not a one-off

The AECJ letter is not unique. It is the current best example of a category of sourcing signal that already includes the 2018 Google walkout organizer list, the Microsoft ZeniMax QA union, the No Tech for Apartheid signatory list, and the 2024 Amazon RTO-mandate NLRB case. Every one of them is a public, named, self-identifying list of dissatisfied insiders at a specific employer on a specific issue.

The macro backdrop makes the category more valuable, not less. More than 92,000 tech workers have lost their jobs so far in 2026 while GPU orders and data center construction keep climbing. Seattle is the largest U.S. city to impose a data center moratorium, one of 54+ local freezes adopted in 2026, with moratorium bills filed in at least 12 states. Each hearing produces a new named-testimony roster. Each roster is a sourcing artifact.

54+

If your sourcing stack does not have a way to ingest dissent artifacts, you are systematically leaving the highest-intent passive candidates on the floor. That is the workflow Refolk handles: point it at a public roster, describe the fit, and get the intersect with your stack, your city, and your level.

FAQ

Is it appropriate to source from an open letter like the AECJ list?

Yes, with two caveats. Signers put their names on a public document intending it to be read. Reaching out about a role, especially one aligned with why they signed, is legitimate. What is not appropriate is scraping the list into a mass sequence, or using the signature as leverage ("I saw you signed, here is a job before Amazon fires you"). Reference the letter as context, not as a threat, and stay narrow on fit.

Why is this better than filtering LinkedIn for "ex-Amazon"?

Because ex-Amazon is inferred preference and AECJ is revealed preference. Someone who left Amazon two years ago may have loved the company. Someone who signed a public letter last month has stated their disagreement by name, with title and team attached. In Refolk's index, the "Amazon Seattle + AI critic" keyword pool is 36 profiles. The letter has 1,000+. You are choosing between a keyword proxy and the actual signal.

How do I find the team affiliations if the letter page is long?

The AECJ letter publishes role and team for each signer once they cross 1,000 signatures. You can read it directly, but if you want to slice by team, level, and location, use a tool that resolves the roster against public professional profiles. Refolk does this natively: ask in plain English for "AECJ signers on AWS Batch or Data Center Automation, SDE III or above," and it will return a ranked shortlist.

What about the 2,400 solidarity signers from other companies?

Those are your adjacent pool. They self-selected on the same axis as the Amazon signers: cautious about "warp-speed" AI deployment inside big tech. If you are recruiting from big tech into a smaller shop with a more deliberate AI posture, the solidarity list is a warmer starting point than a cold Google or Apple search. Same playbook: narrow by team, reference the signature, skip the mass outreach.