Cited · honest · opportunity-framed
This brief does five things, in order: (1) sets out OpenAI's own stated view on AI and the labor transition, quoted and cited; (2) names the gap: no dignified, funded, proof-verified human-work destination at scale; (3) shows where EcoWealth's Work Protocol plugs in, as the three-ladder on-ramp; (4) offers a light, honest agent-readiness note on OpenAI's public surface, credit and gaps, with an evidence log; and (5) makes the ask.
Every quote below is from OpenAI's or Sam Altman's own public writing. The pattern is consistent across five years: work will change profoundly, some jobs will go, new kinds of human work will emerge, and everyone, not just a few, should benefit. That is the exact premise EcoWealth is built to serve.
"OpenAI's mission is to ensure that artificial general intelligence benefits all of humanity," where AGI means "highly autonomous systems that outperform humans at most economically valuable work."
Why it matters: OpenAI defines its own goal in terms of economically valuable work. If the systems outperform humans at that work, "what people do instead" is not a side question: it is inside the mission.
Source: openai.com/charter · openai.com/aboutOpenAI commits "to use any influence we obtain over AGI's deployment to ensure it is used for the benefit of all, and to avoid enabling uses of AI or AGI that … unduly concentrate power," with a "primary fiduciary duty … to humanity."
Why it matters: "Broadly Distributed Benefits" is a named Charter principle. A funded, proof-verified way for ordinary people to earn from real work is one of the most concrete forms that principle can take.
Source: openai.com/charter (principle: Broadly Distributed Benefits)"The price of many kinds of labor … will fall toward zero once sufficiently powerful AI 'joins the workforce.'" … "Even more power will shift from labor to capital."
Why it matters: Altman's own answer was redistribution (the American Equity Fund: a 2.5% tax funding ~$13,500/adult/year). EcoWealth adds the missing half: not just income, but work to do, earned, verified, and dignified.
Source: moores.samaltman.com (March 16, 2021)Altman's OpenResearch gave ~3,000 people $1,000/month for three years. Recipients spent on basics and on helping others, and remained oriented toward work, staying more likely to be searching for a job.
Why it matters: Even with a cash floor, people still reach for work. Income alone doesn't supply the structure, dignity, and progression a person wants. EcoWealth supplies exactly that: a ladder, not just a check.
Source: openresearchlab.org/projects/unconditional-cash-study (NBER papers, 2024)"There will be very hard parts like whole classes of jobs going away … If history is any guide, we will figure out new things to do and new things to want."
Why it matters: "New things to do" is the whole thesis. EcoWealth is a concrete, ready set of new things to do, that also happen to heal the built and natural world.
Source: blog.samaltman.com/the-gentle-singularity (June 10, 2025)"US labor markets grow by adding new kinds of work … the majority of Americans now work in occupations that did not exist in 1940." Who OpenAI is building for: "frontline retail and service teams; skilled trades and field operations; manufacturing and supply chain roles …"
Why it matters: OpenAI itself names skilled trades and field operations, hands-and-feet work, as who it's building for, and even models the ~14,000 construction jobs behind each AI data-center campus. EcoWealth's Build and Restore ladders are that work, made fundable and verifiable.
Source: cdn.openai.com/global-affairs/…/jobs-in-the-intelligence-age-sept-2025.pdf"New jobs will be created; others will evolve. And some jobs will disappear, we should be clear-eyed about that … make sure that everyone, not just a few, benefits." "Jobs are more than a set of static tasks. They require collaboration, judgment and relationships."
Why it matters: OpenAI's own blueprint says the win condition is that everyone benefits, and that real jobs are about judgment and relationships, the human, on-the-ground work EcoWealth funds and proves.
Source: cdn.openai.com/global-affairs/…/oai-workforce-blueprint-oct-2025.pdfOpenAI is building the OpenAI Jobs Platform (matching, launching mid-2026) and OpenAI Certifications (goal: certify 10 million Americans by 2030). That is the credential-and-matching side of the transition. EcoWealth is the doing-labor destination side: the real Build / Grow & Cook / Restore work a certified or displaced person can actually be routed into, funded and proof-verified. One matches and credentials; the other supplies dignified work with a receipt. They fit.
OpenAI names the categories of new human work. Governments and researchers debate the income floor. But between "here is a category" and "here is a check," one thing is missing at scale, and it is the thing people actually want.
A displaced or searching person today can find: job boards (postings, not work), gig apps (narrow, extractive, no progression), training (a certificate, then a cliff), or cash pilots (money, but no work to do). What almost nobody offers is a single place where you can start real work at step one today, be paid when a buyer funds it or credited when not, prove you did it (photo + GPS + method), get a settled, cited receipt, and climb a real competence ladder toward mastery, all while the work itself restores the built and natural world.
That destination is the missing piece under every OpenAI quote above. "New things to do" needs a place to do them, with proof and pay. Building that place, and proving it on-chain, is exactly what EcoWealth has already done.
The Ecological Work Protocol is live on Base mainnet. It turns a place-and-a-problem into a bounded, funded work packet, routes it to a real person, verifies proof before money moves, settles on-chain, and emits a tCO2e receipt where the work is ecological. The human entry point is three competence ladders: each grounded in a respected standard, each starting at a rung anyone can do today.
The packet lifecycle: how one rung becomes a settled, dignified job:
The Kit in document 03 exposes each of these five steps as an MCP (Model Context Protocol) tool an assistant can call, so ChatGPT doesn't just describe the work; it can route a willing user straight into it. That is the plug.
No jargon. OpenAI's security and crawler hygiene are exemplary; the note below is only that the newer agent-legibility conventions (the same ones OpenAI's own products read) aren't on the front door yet. Here is what publishing them is worth.
Your marketing surface publishes no structured brief, so an assistant describing OpenAI works from scraped prose; a single file (the same convention your own products read) lets the company define how it's represented.
There's nothing at the well-known agent-card path, so an agent that wants to reach OpenAI programmatically has no advertised door; publishing one turns OpenAI from readable into reachable by the agent ecosystem it helped create.
No schema.org handshake surfaces to a plain client, so the global-affairs reports an assistant might cite are read as loose prose; structured data means they're quoted accurately.
This is a light, respectful read of what a normal visitor's assistant sees on OpenAI's public front door: passive only, from standard HTTP responses on 2026-07-10. The point is not to grade OpenAI; it's to note that the newer agent-legibility conventions, the same patterns OpenAI's own products consume, aren't yet published on the marketing surface. Credit first, then the gaps.
openai.com/robots.txt = 200, minimal, points to a well-structured sitemap.xml (200) with ~33 sections, including /global-affairs/, where the jobs reports live./.well-known/security.txt (200) with a Bugcrowd program and coordinated-disclosure policy.chatgpt.com/robots.txt explicitly manages crawlers (CCBot, Google-Extended, PerplexityBot, and others), they clearly think about machine traffic.Permissions-Policy, X-Content-Type-Options: nosniff, X-Frame-Options: SAMEORIGIN.llms.txt on openai.com or chatgpt.com: the structured brief that lets an assistant summarize an org accurately isn't published on the front door./.well-known/agent-card.json (404): nothing advertises a callable concierge to other agents.schema.org handshake surfaced to a plain client on the marketing pages we could read.cf-mitigated: challenge), appropriate anti-abuse, noted as OpenAI's own infrastructure and not routed around; it does mean a generic agent driving the page is challenged.None of this is a knock: it's the natural next step for the company whose products read these conventions. And it's exactly the layer the Agent Kit (document 03) provides for the EcoWealth side of a pilot, so a ChatGPT flow has a machine-readable place to route a user.
Evidence log: passive probes, 2026-07-10
| Probe (GET) | Status | What we saw |
|---|---|---|
| openai.com/ (headers) | 403 challenge | Cloudflare bot challenge; HSTS preload + COOP/COEP/CORP + locked Permissions-Policy present. Their infra, not bypassed. |
| openai.com/robots.txt | 200 | 98 bytes; Allow: /, one disallow, links sitemap. Clean. |
| openai.com/sitemap.xml | 200 | Sitemap index, ~33 sections incl. /global-affairs/. Well-organized. |
| openai.com/.well-known/security.txt | 200 | PGP-signed; Bugcrowd + disclosure policy + encryption key. Exemplary. |
| chatgpt.com/robots.txt | 200 | ~3.9 KB; explicit rules for many AI crawlers. Thoughtful. |
| openai.com/llms.txt | not present | Returns the challenge page, not a file. Gap / opportunity. |
| openai.com/.well-known/agent-card.json | 404 | No A2A agent card. Gap / opportunity. |
| openai.com/.well-known/ai-plugin.json | 404 | No legacy plugin manifest on marketing domain (expected). |
All rows are standard public GETs a browser makes. No authentication, no fuzzing, no endpoint probing, no wall routed around.
Nothing here needs OpenAI system access to start. The pilot runs on EcoWealth infrastructure; the Kit is vendor-neutral and portable.
A conversation. Fifteen minutes on where the OpenAI Jobs Platform / Certifications track meets a real doing-labor destination: the complement to matching and credentials.
A single pilot. Let a ChatGPT-driven flow route one willing user into one EcoWealth ladder: step one, funded or credited, proof-verified, settled with a receipt. One person, one dignified job, one on-chain proof.
The Kit, live over EcoWealth's public data. An assistant answers "where can I do real work near me?" with a routable ladder step and a settled-receipt example: no OpenAI systems touched.
✦The through-line
"Whole classes of jobs going away" and "new things to do" are the same sentence with a gap in the middle. That gap is a dignified, funded, proof-verified place to do the next kind of work. It's live on Base mainnet today, waiting to be routed into. See it run in OpenAI's own brand, then hand a person their first rung.
