David Solomon

Goldman Sachs has appeared at three structurally significant altitudes in the AI economy inside three weeks. On May fourth the firm anchored a $1.5 billion vehicle with Anthropic, Blackstone, and Hellman & Friedman to embed engineers and deploy Claude inside private-equity-backed portfolio companies — one hundred fifty million dollars committed. On May thirteenth you boarded a plane to Beijing with sixteen American executives to negotiate trade, artificial intelligence, rare earths, and Taiwan. And the firm continues to lead the underwriting on Anthropic’s IPO alongside JPMorgan and Morgan Stanley, with FutureSearch’s median forecast at $560 billion in March 2027. Three altitudes — operator-base capital, international diplomatic representation, public-markets gatekeeping — three weeks. The firm is placed everywhere the AI economy is being capitalized except one place. I am writing on behalf of the credential body that operates at the one place the firm is not yet placed.

The platform that owns the methodology is Anti-A Industries, Inc. — a Delaware C corporation formed last month, ten million shares of Common Stock authorized at par, registered agent in Lewes, EIN obtained the same day. The Foundation, a Wyoming 501(c)(3), licenses the methodology through a service agreement called the Hinge and operates the credential — SmithFellow Core, a $2,000 in-person on-ramp, twenty-four to thirty hours, graded against a ten-dimension Facilitator Observation Rubric across three checkpoints, taught at five-station facilities in leased space. The facility replicates. The replication target is three thousand locations nationally, structured as six hundred regional networks of five specialized campuses, with Anti-A holding all facility real estate as a Qualified Opportunity Fund. The valuation thesis at 2027 execution rests on conservative reference to the comparable AI-economy transaction landscape — Wiz to Google at thirty-two billion, LinkedIn to Microsoft at twenty-six, Slack to Salesforce at twenty-seven — and on a single structural fact: the workforce side of the AI transition has not yet been built, and the institution that builds it first becomes infrastructure for everything downstream. Station Three is the first facility in the country designed not to teach people how to use AI but to prepare them for what AI becomes — the methodology, SmithTalk, was developed across hundreds of working sessions and operates on three tiers of human readiness: Transactional, Informed, Dialogic.

Your structural counterparty in the May fourth vehicle has personally endowed a $600 million credential body at Tsinghua — a different population, an identical architectural pattern. The firm that wrote the systemic-risk vocabulary for AI workforce displacement in successive annual letters to CEOs was on the same delegation but is not in the vehicle. The convergence is structural. Same coordinated capital, same AI bets, two firms inside the JV, two altitudes nine days apart, and a third firm in the diagnosis but not in the vehicle that operates inside it. What none of these moves address — what cannot be addressed inside any vehicle whose integration thesis depends on engineers embedded in portfolio companies that have already capitalized themselves — is the workforce base those portfolio companies will hire from. That layer has not been built. That layer is what CrowdSmith was constructed to build.

The credential body was built by Robb Deignan. He has thirty-seven invention concepts active in a standalone pipeline — inventions he developed during the twenty years he could not afford a patent attorney, evaluated through a proprietary methodology called SmithScore that scores concepts on operational, market, and feasibility dimensions; surviving concepts are validated through SmithForge and documented in the Patent Ledger. The pipeline runs through Station Five, where prototyping produces robot-demonstrated manufacturing proof for inventor concepts. He built the credential body he wished had existed for him. He developed the methodology across hundreds of working sessions with the platform that runs Station Three, from a kitchen table in Tacoma, on Social Security Disability Income. He is sixty years old. He was living on his own at sixteen.

Daniel Nachmann, Anthropic’s president, said publicly at the JV announcement: “Having the model alone doesn’t change your workflows or how you operate. You need people who can combine the technology with what’s actually happening in the business and implement those changes.” That sentence is the diagnosis. The vehicle is the operator-side answer to it. The workforce-base side has the same diagnosis, a different population, and the same gap. The architecture pattern is the one Goldman has underwritten at scale before — institutional credential bodies whose graduates become the labor base of an emerging economy. The MBA infrastructure the firm itself recruits from is one example. The recognition CrowdSmith asks Goldman to take is that the workforce credential for the AI economy is the institutional category the firm has placed capital, capital-markets representation, and international diplomatic presence around — but has not yet placed in.

The firm appears at three altitudes in the AI economy in three weeks. The workforce-base altitude is the one absent from all three appearances. Whether the absence is by design or by oversight is the question this letter opens. The architecture is built. The credential body is operational. The methodology is on the public site. The Convergence is on the record. The conversation about whether Goldman is the firm that capitalizes the workforce-base half of the integration architecture — and at what scale, on what timeline, in what structure — is the conversation that has not yet been had with anyone.