Your AI Knows You. Egoist Machines Wants the Passport

In partnership with

You Already Have a Take on What AI Does Next

OpenAI or Anthropic? Which model leads the next benchmark? Which company ships the next major breakthrough?

If you follow AI closely, you already have opinions on where the industry is headed. Kalshi lets you trade on real-world AI and technology events, with markets that move as models launch, benchmarks drop, and announcements happen.

The people who follow this space most closely often see the story developing before everyone else. Put that knowledge to work and trade what you think happens next.

Trade $25, Get Up to $500

_{Bonus credit varies from $15 to $500. Terms apply.}

At two in the morning, the modern AI user is a minor bureaucrat in a republic of amnesiacs.

You tell one machine that you hate red-eye flights, another that your left knee is unreliable, a third that you are allergic to walnuts, and a fourth—God help you—that your preferred tone is “warm but concise.” Then you open a new app and begin the ceremony again. Name. Taste. Work. Goals. The private liturgy of who you are, shoveled into another glowing box.

Egoist Machines, a two-person Y Combinator company from the Summer 2026 batch, proposes a document for crossing these borders: the AI Passport, a user-owned memory layer meant to travel between AI products. Your preferences, history, goals, purchases, relationships, and identity sit in one structured record. An app requests specific fields for a declared purpose and duration. You approve, edit, or deny the request. If the app learns something durable, it proposes a write. Every read and proposed update leaves a receipt.

It is OAuth for the self, with a better wardrobe and a much larger blast radius.

The pitch: stop reintroducing yourself to machines

Egoist’s consumer promise is clean enough to write on a customs form: carry your context across AI applications without starting over. The website says the personal product will be private by default and free for life; as of this profile, it is inviting users to a waitlist.

The developer pitch is more ambitious. Businesses can ask for narrowly scoped context—clothing size, budget, style preferences—then personalize an e-commerce experience from the first click. A support system can retrieve prior issues and account history so each new agent does not interrogate the customer from zero. The company describes a permissioned read/write loop in which applications request exact fields, users control access, and receipts make the exchange inspectable.

That matters because today’s memory is mostly trapped inside products. ChatGPT maintains saved memories and chat-history context within ChatGPT. Other assistants and applications are building their own islands. The inconvenience is obvious; the strategic consequence is nastier. The product that owns your history accumulates switching costs. “Personalization” becomes a velvet-lined lock-in mechanism.

Egoist wants to pry the memory loose from the app and hand it back to the person.

The machine under the passport

The public materials show more than a mood board. Egoist’s developer page describes scoped field requests with purpose and duration, user approval, context retrieval, proposed memory writes, and an audit receipt. It advertises a unified layer spanning lifestyle services, AI assistants, utilities, and knowledge tools, plus “800+ other connector sources.” Its privacy policy says connected health or fitness data is consent-gated, that memories are not used to train models, and that Google user data is accessed read-only and on demand rather than stored.

Those are company statements, not an independent security audit. Still, they reveal the intended architecture: Egoist is not merely exporting a text file called things_about_me.txt. It is trying to become a broker for structured personal context, permissions, provenance, and revocation.

There is also a public-facing identity card at ego.ist/i/you, pitched as a free Linktree-like page where people can collect links, interests, and things sent by others. This is clever distribution. A public profile attracts ordinary humans; the private memory graph grows behind it; the developer layer gives apps a reason to ask for access. The cute card is the bait shop. The real business is the canal.

The founders: security brain, attention engine

David Khachaturov brings the kind of résumé that makes a permission system feel less like cosplay. His YC bio lists a BA/MA and PhD in computer science from Cambridge, an MSc from Oxford, and says he became the youngest-ever Fellow at St Catharine’s College at 25. It also lists previous work as an OpenAI machine-learning researcher on PaperBench, lead AI security researcher at FrontierOne, Cambridge computer-science teaching across eight-plus courses, and a prior game-publishing company with a ten-title portfolio.

Erin McGurk brings a different form of infrastructure: attention. Her YC bio says she built a following of more than one million across social platforms under @erinmerylstudy, with interests spanning AI safety, privacy’s social and legal infrastructure, space governance, urban design, and advertising psychology. Her own description ends with a joke about being very good at talking and getting people to listen. In an industry where distribution is often the corpse hidden under the product roadmap, that is not a minor asset.

YC lists LinkedIn profiles for both founders:

• David Khachaturov: https://www.linkedin.com/in/david-g-khachaturov/

• Erin McGurk: https://www.linkedin.com/in/erin-mcgurk

YC does not list a Twitter/X link for either founder in the company bio. The company’s YC primary partner is Harshita Arora.

The competitive weather

Egoist is entering a storm, not an empty field.

The large AI platforms can extend their own memory systems and make portability unnecessary for users who never leave. OpenAI’s own documentation already gives users controls to inspect and delete saved memories, while tying those memories to the ChatGPT account. Developer-memory companies such as Mem0 sell infrastructure for applications to remember users. Customer-data platforms, identity providers, password managers, personal-data stores, MCP connectors, and old-fashioned CRM vendors all occupy pieces of the territory.

Egoist’s wedge is not memory alone. It is cross-application memory governed by the user: a neutral passport instead of a pile of corporate dossiers. That neutrality is its differentiation and its structural headache.

To work, the passport needs two groups to arrive at once. Users need enough connected context to make it valuable. Developers need enough passport holders to justify integration. Large platforms may resist a neutral layer that weakens their lock-in. Small developers may love the concept but hesitate to depend on a tiny startup for identity-adjacent infrastructure. Every additional data source increases usefulness and expands the attack surface.

The company must also solve the semantic problem: human identity is not a tidy database. Preferences decay. Relationships change. A “fact” inferred by one model may be a hallucination wearing a name badge. Consent screens become wallpaper. Revocation is only meaningful if downstream systems honor it. A receipt proves that access occurred; it does not make the recipient forget.

And then there is the central paradox. To protect users from fragmented corporate memory, Egoist proposes concentrating an extraordinarily intimate map in a new intermediary. The passport office becomes one of the juiciest targets in the city.

Traction: the honest blank space

Egoist’s public YC profile lists a two-person team, active status, and a San Francisco location. Its site has a consumer waitlist and a developer-access form. The company claims support for more than 800 connector sources and presents detailed product flows, but it does not publicly disclose revenue, active users, customer names, funding beyond YC backing, or independently verified usage metrics.

That absence is not a scandal; this is a newborn company in a current YC batch. But it is the line between a vivid thesis and a proven business. The passport is currently more border architecture than bustling nation.

Why it matters

If AI becomes the interface to everything, memory becomes a control plane. Whoever decides what the machine knows about you can influence what it recommends, what it hides, what it charges, and which version of you arrives in the next room.

Egoist Machines is making a sharp bet: people will demand portability and granular consent before the dominant platforms make those ideas irrelevant or cosmetically imitate them. The company’s best insight is that AI memory is not merely a convenience feature. It is identity infrastructure.

The opportunity is enormous because the current experience is absurd. The risk is enormous because the cure requires a new keeper of the keys.

Somewhere in the neon rain, another app is asking who you are. Egoist wants you to flash one card, reveal exactly enough, collect the receipt, and keep walking. That is a civilized vision. It may also be the beginning of the most consequential passport regime nobody voted for.