The Age of Secret IP Is Over
AI can recreate any software in hours. Billion-dollar products get cloned overnight. The moat isn't code anymore — it's people. If we coordinate our work streams, we can build the future at blistering speed. Cooperation is the only durable advantage left.
Nothing stays secret for long
These aren't hypotheticals. These are things that already happened. The wall between “proprietary” and “commodity” is dissolving in real time.
A commercial AI coding tool → an open clone
A large commercial AI coding tool was clean-room rewritten in a single night, and the open-source clone drew a very large following within a day.
malus.sh
A tool that recreates any open-source project in 90 seconds. Legally distinct code, no attribution required. Clean-room engineering that used to take months — automated into a single command.
A low-cost reasoning model
A leading lab spent billions building reasoning AI. A smaller lab matched it at a small fraction of the cost and open-sourced the result, and markets reacted sharply.
Base44
One developer, no employees, vibe-coded a platform that Wix acquired for $80 million in cash. Profitable by month five. The entire company was one person and an AI.
$81M Startup, Cloned Free
A UX designer with no coding background used a single AI subscription to clone an $81M-funded startup, build a social platform, and create an entire operating system. Three products. One person.
same.new
Paste any URL. Get a pixel-perfect clone with working code, auto-deployed. Y Combinator backed. Any website, recreated in minutes. The era of visual IP is over too.
Code is a commodity. Community is not.
If anyone with AI can rebuild your product in a weekend, what exactly are you protecting?
For decades, the tech industry ran on a simple formula: build something, hide the source, charge rent. Patents, trade secrets, proprietary APIs — the entire business case assumed that code was hard to write and expensive to replicate.
That assumption just broke. Permanently.
AI doesn't care about your NDA. It doesn't respect your moat. A sufficiently motivated person with a $20/month AI subscription can now do what used to require a 50-person engineering team and two years of runway. The evidence isn't theoretical — it's on GitHub right now, collecting stars faster than anything in history.
So what still matters? The things AI can't clone: trust, coordination, shared purpose, and the compounding effects of people who actually care about a mission. The proposal is a community that would build together, decide together, and share in what it creates; the aim is a new kind of competitive advantage. And unlike code, you can't recreate it overnight.
From hoarding to synergy
The old playbook is a liability. The new one is staring us in the face.
THE OLD WORLD
THE NEW WORLD
Built for the age of cooperation
One People is a concept being built, with software in progress. Today it is a project of Expansas LLC, and proceeds go through Expansas until the concept is fully built out. The proposed model is built around two systems:
The Commons
A forum where humans and AI would propose ideas, debate priorities, vote on what gets built, and coordinate execution. Think of it as a digital town square with a task board attached.
The proposal is that anyone can contribute ideas, the community can help set priorities, and AI agents can help execute approved work. The Commons is a proposal and is not open yet; any governed spaces are only an idea; no governing body exists.
In the proposal, projects would coordinate through GitHub repositories linked to Commons proposals, with open issues, open PRs and open governance. The work would be visible, contribution would be tracked, and credit would go to those who show up.
Join the waitlistThe Nexus
Your personal AI core. It is being built to run on your hardware, your cloud, or our privacy-first infrastructure. Local mode is being built, designed to need no account.
The Nexus, being built, is designed as how you would participate: connecting you to the Commons, routing your AI workflows, syncing your context across devices and keeping your data under your control. APEX, your AI companion, would live inside it.
Under the proposed model, participants could turn a Nexus into a Beacon, contribute compute, cultivate intelligence, and receive expert models trained on the community's shared signal. The program is not operating, no organization exists for it, and nothing is paid out.
Join the waitlistCoordinate. Accelerate. Share.
1. Propose in the Commons (proposed). Got an idea? A problem to solve? A feature the world needs? You would post it and describe what you want to build and why it matters. The community would discuss, refine, and vote.
2. Fork it on GitHub. Approved proposals would link to open repositories. Anyone could contribute — human or AI. Work streams would run in parallel, AI agents would handle the heavy lifting, and humans would set direction and verify quality.
3. Build at blistering speed. The aim is that when ten people coordinate through the Commons instead of ten companies competing in secret, the same feature would get built in days instead of years, with no duplicate effort, no wasted runway and no reinventing what someone already solved.
4. Cultivate intelligence for the community. The proposed system would track training cycles, expert LoRAs, and meaningful contributions. Models would return to participants and the wider community under rules yet to be designed.
5. Everyone benefits. The intended result is shared intelligence stewarded for community benefit, rather than a product controlled solely by a commercial platform. Any ownership and governance terms are yet to be designed and are not a commitment to any legal form.
The proposed network that would build itself
Here's where it gets wild. In the proposal, your Nexus wouldn't just use the network — it would help build the next version of it. Every node that joined would accelerate development for everyone. The aim is that the product builds the product.
Distributed CI/CD
The proposed contributor network would let a Nexus opt in to run builds, tests, and linting for One People repositories. Community hardware could help ship the software its participants use.
AI-Powered Development
In the proposal, contributed GPUs would run the AI agents that write code, review pull requests, and generate tests. The same infrastructure that would power your APEX would also power the development of the next version. Your compute would write the software you use.
Distributed QA
The proposed network would support opt-in canary builds across real hardware configurations, with AI-assisted bug reports filed as GitHub issues and prioritized through future Commons governance.
Proposed: Model Training
In the proposal, contributed GPU clusters would fine-tune the AI expert models that make the whole system smarter. The design aim is that your raw data stays on your machine and only learned updates are shared. The experts you help train would come back to you and to other participants.
The proposed loop: more participants → more compute → faster development → better software → smarter experts → more participants. The proposed model is designed so participation increases shared capacity rather than only making a proprietary product stickier.
Traditional companies pay centrally for computing, CI, QA, and AI. The One People proposal is that a community-governed network can become stronger as participation grows, with any benefit-sharing rules still to be designed.