Amazon's 1,200-Signer AI Letter Is a Flight-Risk Directory
Amazon Employees for Climate Justice just handed sourcers a roster of 1,200 self-identified flight risks. Here is how to work it before May 2026.
On January 28, 2026, Amazon confirmed 16,000 more corporate cuts, taking the total to 30,000 since October 2025. In the same week, Amazon Employees for Climate Justice (AECJ) published an open letter signed by more than 1,200 current Amazon employees demanding worker input on AI-driven layoffs. If you source engineers, that letter is not a news story. It is a directory.
Why a protest letter is the cleanest flight-risk list in FAANG
The AECJ letter is a self-identified, publicly attributable roster of 1,200+ current Amazon employees who have already told their CEO they disagree with strategy, during the same quarter the company is executing the largest workforce reduction in its history. That is a sourcing artifact you almost never get from inside a FAANG.
Compare it to what you normally work with:
- LinkedIn "Open to Work" is noisy and lagging, and Amazon's PIP-heavy culture teaches people not to flip it on while employed.
- Blind posts are anonymous, so you can read sentiment but you cannot reach anyone.
- WARN filings give you counts and cities, not names.
- Internal referrals require a warm intro you probably do not have.
AECJ, by design, publishes signatory job titles and orgs once a letter clears 1,000 signatures. This one cleared that threshold, so the page is functionally a directory of principal engineers, senior PMs, applied scientists, economists, and marketing leaders who have publicly antagonized Andy Jassy on the record. You do not have to guess who is unhappy. They wrote it down.
What the numbers actually say
The 1,200 signers are roughly 0.34% of Amazon's ~350,000 corporate headcount, and about 4.0% of the 30,000 roles being eliminated. That tiny share is the whole point: this is not mass sentiment, it is a filter for the specific subset with the conviction, tenure, and outside options to publicly criticize their own CEO.
| Slice | Count | Source |
|---|---|---|
| AECJ letter signatories, current Amazon | 1,200+ | GeekWire, Jan 28 2026 |
| AECJ external solidarity signers | 4,034 | amazonclimatejustice.org/solidarity |
| Amazon corporate cuts, Oct 2025 to May 2026 | 30,000 | Amazon, Jan 28 2026 |
| Prior AECJ climate letter peak | 8,700 | NBC News |
| Amazon "climate/sustainability" pros surfaced, US | 26 (18 still at Amazon) | Refolk's index |
| Amazon/AWS senior PMs surfaced, US | 103 | Refolk's index |
| SDE/Principal Eng/Applied Scientist matches, Amazon Seattle | 14 (11 at Amazon, 10 in Seattle) | Refolk's index |
| Signatory share of Amazon corporate | ~0.34% | Derived |
| Signatory share of 30k affected roles | ~4.0% | Derived |
The last three rows are the ones to stare at. The universe of Amazonians who publicly identify with climate or sustainability work is astonishingly small, and it overlaps almost perfectly with AECJ's organizing core. For a climate-tech buyer, this is not a flight-risk list. It is a total addressable market.
The five-month sourcing runway most recruiters will miss
Amazon's layoffs are phased through WARN notices in Washington and California running January 26 through the end of May 2026, which means the flight-risk cohort refreshes weekly, not once. If you treat this as a single event and pull one shortlist in February, you will miss roughly 80% of the runway.
The mechanism is straightforward. WARN notices land in waves as Amazon confirms which orgs (AWS, Alexa, Prime Video, devices, advertising, last-mile delivery) are absorbing which share of the 16,000 January cuts and the roughly 14,000 already taken since October. Each wave adds a new tranche of newly-severed or newly-nervous employees to LinkedIn, GitHub, and Blind. Each wave also shifts the internal calculus for signers who were on the fence in January and are watching their skip-level disappear in March.
The correct posture:
- Build a saved query against the AECJ signatory titles and rerun it every Monday for 20 weeks.
- Cross-reference each new WARN filing against the org names AECJ publishes.
- Segment your outreach by wave, not by list, so you are not re-pinging the same senior PM every three weeks.
- Track the "Project Dawn" codename (the accidentally-leaked internal name for the January plan, sent to AWS employees by mistake) as a boolean anchor on Blind, Hacker News, and public Discord archives.
This is the exact work that breaks in a spreadsheet. Describing the slice in plain English to Refolk ("Amazon Seattle principal engineers or senior PMs, still employed, publicly signed the AECJ letter or posted about Project Dawn in the last 30 days") is faster than maintaining seven boolean strings across three tools.
The outreach wedge Jassy handed you
Andy Jassy publicly attributed the October cuts to "culture," not AI or cost. The AECJ signatories publicly said the opposite. That mismatch is your first line.
On October 31, 2025, Jassy told GeekWire the layoffs were about agility, not AI or cost-cutting. Five weeks later, 1,200 of his own employees signed a letter saying AI is exactly what is driving the cuts and demanding worker input on it. Every signer on that letter has already told you they do not accept the official framing. That is a gift.
A cold message that opens with "Saw you signed the AECJ letter. Your CEO said the cuts were about culture and you said they were about AI. Curious which framing is holding up inside your org this week" will out-perform any templated "exciting opportunity" note by a wide margin, because it is specific, it is respectful of what the person actually did, and it invites disagreement rather than a pitch.
Every signer on that letter has already told you they do not accept the official framing. That is a gift.
Segmenting the 1,200 into pools you can actually work
The 1,200 signers are not one audience, they are at least four, and the outreach that works for a principal engineer will annoy a Whole Foods cashier. Split the list before you write a single message.
Pool 1: The 26-person sustainability core
Reported signatory composition spans senior engineers, product leaders, data scientists, economists, marketing managers, Whole Foods cashiers, IT support, and warehouse associates. Inside that, the sustainability-identified subpopulation is tiny. In Refolk's index, only 26 US professionals surface for Amazon plus climate/sustainability keyword combinations, 18 of them still at Amazon and 8 in Seattle. That is your entire market for a Watershed, Patch, Stripe Climate, Sila, or Form Energy search. Work it end-to-end in a week.
Pool 2: Senior PMs and principal engineers
103 US Amazon/AWS senior PMs surface in Refolk's index, and Principal PMs dominate the top of that ladder (15 of 25 in a recent sample). This is the exact profile most cited among AECJ signatories and the exact profile Anthropic, Hugging Face, and every mid-stage AI infra company is trying to hire this quarter. Fourteen SDE, Principal Engineer, and Applied Scientist matches surface on "Amazon Seattle" alone in the same index, with 11 still at Amazon and 10 in Seattle.
Pool 3: The 4,034 external solidarity signers
The AECJ solidarity page shows 4,034 external supporters as of retrieval, including workers from Google, Apple, Meta, and Microsoft. These people have publicly aligned themselves with an anti-AI-rollout stance and cluster geographically in Seattle and the Bay Area. For an AI-safety-forward employer, this is a secondary but larger pool with the same personality filter and lower switching friction than an Amazonian mid-layoff.
Pool 4: The AECJ alumni network
Eliza Pan, a former Amazon program manager who left in 2019 and stayed involved with AECJ after the NLRB retaliation cases, is the template. Prior AECJ organizing peaked at 8,700 signatures on an earlier climate letter, and many of those signers have since scattered to climate tech, policy shops, and other FAANGs. They are pre-warmed intros to current signers. Find them, and your reply rate on the current 1,200 climbs.
Destination employers who should already be working this list
If you recruit for AI safety, climate tech, or worker-first infra, the AECJ letter is a hiring plan handed to you free. Named destinations that map cleanly to AECJ's three public demands (renewables, worker input on AI, no AI for surveillance or deportation):
- Anthropic, Hugging Face: AI-safety framing that maps directly to the letter's AI critique.
- Watershed, Patch, Stripe Climate: climate-first positioning for the 26-person sustainability core.
- Recursion, Sila, Form Energy: applied-science and hard-tech shops that need Amazon-scale eng and PM talent aligned to a mission the candidate has already publicly endorsed.
For any of these employers, the pitch writes itself and the addressable universe is small enough to work by hand. What is hard is keeping the list fresh as WARN waves land through May.
What to build this week
The minimum viable version of this playbook is three saved searches and a Monday reminder.
- Signer directory pull: names, titles, and orgs from amazonclimatejustice.org/open-letter, cross-referenced to LinkedIn and GitHub for current confirmation of employment.
- Weekly WARN diff: new California and Washington filings against Amazon entities, matched to the org buckets AECJ publishes.
- Personality-adjacent expansion: for each confirmed signer you cannot reach, find two co-workers on the same team who liked, reposted, or publicly commented on the letter. That is your warm second ring.
You can build this in a browser and a spreadsheet. It will take a sourcer about a day and a half to stand up and roughly two hours a week to maintain. The deadline is real: the last WARN wave lands in late May, and by June most of the 1,200 will have already talked to someone.
FAQ
Is it legal to source candidates from a public petition?
Yes. AECJ publishes signatory job titles and orgs by design once the letter passes 1,000 signatures, and the letter itself is hosted on a public website with no login. Using publicly posted professional information to identify potential candidates is standard sourcing practice and sits well inside the same legal frame as sourcing from a conference speaker list or a public GitHub profile. The wrinkle to avoid is inferring protected characteristics from the political content of the signature. Reach out on the basis of role and public statement, not inferred belief.
How is this different from just pulling a list of laid-off Amazonians?
The signers self-selected on conviction, not on being cut. Only about 4.0% of the 30,000 affected roles are on the letter, and most signers are still employed. That gives you a five-month runway to build a relationship before they are on the market, rather than joining the queue of recruiters messaging them the day their LinkedIn flips to Open to Work. It is also a much stronger personality filter: signing publicly against your own CEO predicts response rates on cold outreach far above the FAANG baseline.
What outreach angle actually works on AECJ signers?
Lead with the Jassy "culture" quote versus their signature, and ask a real question rather than pitching. Signers have already demonstrated they will engage in a debate about Amazon's AI strategy on the record, so an opener that treats them as a peer with a view (not a resume) will out-perform any templated recruiter note. Save the role pitch for message two, after they have replied. If you are recruiting for Anthropic, Hugging Face, Watershed, or a similarly aligned destination, name the alignment explicitly in the first message.
Will this list be stale by summer?
The signer identities will not be, but their employment status will churn weekly through May 2026 as WARN waves land in Washington and California. Treat the 1,200 as a durable relationship pool and the WARN cross-reference as the freshness layer. The recruiters who win this cohort will be the ones running a weekly refresh from January through June, not a single February pull.
Try it on the search you came here for
Stop building boolean strings. Just describe the person.
Type one sentence. I plan the search, read GitHub, public LinkedIn and Crunchbase records, and the open web as it is right now, and hand back a ranked list with the reason next to every name.
01Describe them
One plain sentence. Role, city, stack, stage, whatever matters to you.
02I read the web live
GitHub, public LinkedIn and Crunchbase records, the open web. Not a database that went stale last quarter.
03You read the shortlist
Ranked, with the reasoning under every name. Open a profile, ask a follow-up, narrow it down.
- Staff backend engineers in NYC who shipped Rust in production
- Series A fintechs in SF under 50 people, growing headcount this year
- Maintainers of fast-growing Rust web frameworks on GitHub
- No boolean, no filters, no seat to buy. One box.
- Read at search time, so a profile updated yesterday counts today.
- Every step visible as it runs, every name with its reason.
500 free credits on sign-up. No card, no demo call. See real searches.