From medium By: Derek Edws, Managing Partner @ Collab+Currency

My mom left the Philippines when she was 21 years old — armed with a degree and $50 in her pocket.

She would navigate America with a quiet determination.

Every educational course became an investment in her professional agency.

Every dollar became a brick in her financial foundation.

She would earn her Master of Nursing at UCLA, and go on to work directly with countless military veterans in VA hospitals over a 40+ year nursing career.

Like most Americans at the time, both naturalized and natural born, her mobilization was made possible by a unique country with three features:

(1) An open economic system;

(2) incentivizing merit-based contributions;

(3) with established property rights to protect ownership over new wealth generated.

Work hard. Outperform. Store wealth.

Compound these actions over a lifetime, and anyone could enjoy a piece of the American dream.

As I expand upon in this essay, I’m no longer confident our legacy playbooks will hold up against the emerging truths of tomorrow — namely, the breathtaking acceleration of artificial intelligence, and its disruptive potential on American labor.

I believe it’s critical for America to start navigating toward a new framework to match new labor realities.

And it starts by understanding the two most disruptive technological innovations of the last twenty years:

AI and Crypto.

Artificial Intelligence

Thirty years ago, internet search engines were introduced to the world.

Their value was deceptively simple:

Retrieve the world’s data.

Twenty five years later, OpenAI’s ChatGPT is the fastest growing consumer application in history. The product enjoys 300M weekly users, with over 10M in monthly paying subscribers. In just under two years, the product generates roughly $3B in annual revenue.

The value of these AI systems?

Combine “retrieval” with “completion”.

Put simply:

(1) Request any job using natural language; and

(2) Outsource the job’s completion to computer intelligence.

Over the last three months, it’s become clear that these products are rapidly improving in their ability to complete these job requests, across a growing number of modalities — text, math, audio, video, geometry, coding.

And so far, the scaling hypothesis for improving AI systems is holding.

More data. More compute. Better models.

Further, new dimensions for scaling these models are being explored, with innovations like test-time compute emerging. Last month, OpenAI’s latest reasoning model completed a 25% score on the frontier math benchmark, composed of problems that only deeply specialized mathematicians can solve.

These next generation reasoning models are taking significant steps adapting to novel tasks with observable answers.

Meanwhile, new advancements in robotics are making machines increasingly capable of performing complex physical tasks.

Intelligent, human-like hardware will:

- Never get tired.

- Never take weekends off.

- Never go on strike.

Technological productivity unlocks like this are not without precedent.

The industrial revolution (18th century) and digital revolution (20th century) radically reduced costs and increased efficiencies across various forms of labor, fundamentally reshaping markets and economic structures.

But, artificial intelligence already feels like a different type of technological revolution.

These systems are able to create value in ways previously thought impossible or impractical — unburdened by legacy cost structures.

Anthropic CEO Dario Amodei believes that AI may give us the next 50–100 years of biological progress over the next 5–10 years, simply by offloading most of our human labor structures to AI-enabled systems.

As the infrastructure to support the growth of these systems continues, I believe two trend lines will become obvious to humans around the world:

- The cost of all intellectual labor will trend down to the cost of running AI systems; and

- The cost of all physical labor will trend down to the cost of a robot’s mechanical parts.

The promise of this technology is clear:

Unbounded outcomes.

Radically cheaper costs.

Across every category of value.

For the purpose of this essay, let’s assume the above trends are true.

As someone who deeply believes America’s next generation should enjoy access to the same opportunities as the generations that came before — there are two questions I can’t shake:

First, how can America win the AI revolution on the merits?

Second, how can individual Americans participate in the upside of tomorrow’s AI systems — recognizing its disruptive potential on human labor?

Enter Crypto

In 2008, Satoshi Nakamoto published the Bitcoin whitepaper.

The paper laid out a new, gamified accounting system where distributed computers all over the world could work together to agree on a shared digital truth — namely, the state of Bitcoin’s ledger.

Today, Bitcoin is the most powerful supercomputer in the world.

The network is orders of magnitude larger than the combined network sizes of Amazon, Google, and Microsoft.

Yet fifteen years in, blockchains are still clunky.

- Poorly written code can create vectors for theft or zeroed-out account balances.

- User mistakes managing private keys led to the largest share of stolen crypto last year.

- Decentralized applications are difficult to use compared to legacy web applications — with astronomically high rates of user churn.

But even with these limitations, interaction with crypto is at an all time high.

One study estimates that 40% of Americans now own cryptocurrency, up from 30% in 2023.

Nearly 24,000 monthly developers around the world actively contribute code to blockchains and blockchain-based applications, up from 1,000 monthly developers ten years ago.

Given all its limitations — why does crypto keep growing?

I believe it is because crypto enjoys five features that — working together — can’t be replicated by any other database architecture on planet earth:

Digital Ownership. Blockchain databases are global, fully auditable, community-owned, tamper-proof, and run 24/7 across the internet.

Using blockchains, individuals can enjoy sovereign ownership over any digital object on the internet — laying down a global digital property rights system for the first time in human history.

Coordination Incentives. Using autonomously executing contracts, blockchain-based protocols can wield programmable incentives to coordinate new types of digital work on each network.

This can include incentives for those who use a product or service, provide economic security, contribute core code, provision marketplace supply or demand, refer others to the product or service, and much more.

Frictionless Micropayments. Today, prohibitively high transaction costs force most internet businesses to adopt business models like subscriptions, bundling, and ad-supported services. This constraint limits business model innovation across the internet, restricting consumer-friendly alternatives that could unlock new product/service revenue.

Conversely, blockchain databases facilitate frictionless, 24/7, low-cost, digital payments on a global scale — bypassing digital payment inefficiencies and enabling low-fee, instant transactions without the risk of chargebacks. Further, any cryptoasset can be divided into an arbitrarily small denomination.

Shared Standards. By leveraging shared settlement standards on top of blockchains, various types of protocol tokens, stablecoins, applications, games, and financial services can all seamlessly compose with one another’s logic — much like how lego blocks can connect regardless of color or size.

Distributed Security. Blockchain networks are often distributed across numerous nodes around the world, eliminating single points of failure. This decentralized architecture makes it far more difficult for malicious actors to compromise these networked systems, as it requires control over a majority of nodes simultaneously.

Today the market cap of the crypto economy is roughly ~$3.6 trillion — across a number of emerging verticals.

And over the next ten years, I believe the crypto economy will be repriced radically higher — spurred largely by the growth of AI’s intersection with crypto across two major trend lines —(1) AI x Crypto Infrastructure and (2) AI x Crypto Applications.

AI x Crypto Infrastructure

To understand the current landscape for AI infrastructure, it’s useful to look back at a historical analog.

In 1849, the California gold rush quickly drove new investments into the region.

Hundreds of new roads were constructed to facilitate rapid transport. San Francisco’s port became one of the busiest in the world — shipping in gold prospectors, goods, and tools from around the world. Robust banking and financial systems were developed to network the demands of an emerging global enterprise.

The infrastructure investments made during this time would create the foundation for the region’s future as an economic powerhouse.

175 years later, the world is now witnessing a similar gold rush — one to create AGI.

This time, however, the infrastructure designed to support artificial intelligence is not confined to a specific region.

These networks — encompassing data, compute, and power — are being constructed by competitors around the world.

Not surprisingly, the capital and computation required to build, train, and optimize AI infrastructure is prohibitively expensive, accessible only to a select few.

Conservative estimates suggest that training GPT3 cost roughly $4M+ for a single training run. For GPT4, estimates suggest the training run cost roughly $60M+.

More capital. More compute. Better performance.

While I am deeply proud and supportive of the work accomplished in America to build artificial intelligence under our traditional corporate forms, I also believe it’s possible to acknowledge their structural limitations:

Captured Value. Although centralized companies wield risk capital to drive meaningful AI innovations, the economic benefits of these products are narrowly scoped to a small group of equity holders, limiting broader societal impact.

Proprietary Knowledge. Advancements in technology frameworks often remain proprietary to centralized firms, slowing overall progress in the field by restricting access to crucial breakthroughs — at a time where information now moves 24/7 at internet speeds.

Opaque Systems. Concentrating synthetic intelligence in opaque, closed, centralized systems makes it impractical for independent verifiers to audit company practices regarding data collection, safety, and accountability.

Closed Competition. The immense computational requirements for advanced AI development create significant barriers to entry for new products to enter, allowing only a few well-funded giants to participate. Opening up competition to a wider range of participants, from individual researchers to smaller teams, may foster a more diverse AI ecosystem — leading to breakthroughs that may be otherwise overlooked.

I believe that by orienting America’s AI infrastructure towards crypto’s five unique features — digital ownership, coordination incentives, frictionless micropayments, shared standards, and distributed security — we can both temper the downsides of centralized AI, and renew the spirit of competition that has long been a hallmark of American capital markets.

Further, by availing America’s AI infrastructure to crypto, I also believe it will lead to AI systems with (a) better performance, (b) greater transparency, and (c) more equitable ownership across millions of American participants in the future.

(a) Better Performance

To understand the breakthroughs that can be achieved in AI without significant funding, look no further than the DeepSeek team. Two weeks ago, the China-based research group released DeepSeek-V3, a 670B parameter model rivaling the performance of many closed-source SOTA models, including GPT-4o and Claude-Sonnet-3.5.

DeepSeek has taken no venture funding to date.

As open source projects like Bitcoin and Ethereum have proved, issuing programmable incentives to a global pool of contributors can supercharge highly qualified labor and compute networks to form on the internet — far greater than what single labs or centralized systems can achieve.

Under this lens, creating systems for rewarding AI labor and compute networks over the internet is no different than creating systems for rewarding Bitcoin labor and compute networks over the internet.

A few examples:

For improved training data, crypto networks can reward human contribution for curating high-quality, labeled datasets — including private data, proprietary intelligence, or information otherwise not found in traditional web scrapes.

For more robust compute networks, crypto networks can incentivize individuals and organizations to contribute tiers of computational power over a decentralized marketplace — bootstrapping a global network of machines on demand without upfront capital investment.

For more performant model training, a network of open source developers can contribute, customize, and improve upon existing models in exchange for tailored rewards. Further, when code and model weights are openly available, hundreds of researchers and developers can simultaneously publish improvements, debug issues, fine-tune custom models and agents, and create novel applications on top — aligned by a network’s incentive model.

Over time, I believe this style of broad collaboration — pioneered by decentralized projects like Nous Research, Prime Intellect, and Bittensor — will outpace what even well-resourced private companies can achieve internally.

(b) Greater Transparency

Open source AI models allow their training process, architecture, and behavior to be thoroughly examined (and improved) by a broader research community. This transparency can help identify potential risks or biases early, translating into more reliable systems that people can trust.

By leveraging blockchains during this process, the entire stack for creating, rewarding, and improving AI protocols can remain transparent and auditable.

(c) More Equitable Ownership

Crypto networks, designed for individual verticals across the AI stack, will establish a more equitable ownership structure than current centralized models for creating synthetic intelligence.

With programmatic incentives, all contributors and participants to a crypto protocol can be transparently compensated for their contributions.

Further, entire markets can form around *any* job in the AI infrastructure stack — leading to granular competition across all categories of AI design. Categories and sub-categories across data, compute, training, and deployment can all compete and accrue value in isolation of one another.

Ultimately, however, it’s not just AI infrastructure that aims to benefit from crypto rails.

I believe AI agents are on track to completely reorganize the current trajectory of global crypto adoption — across every application vertical.

AI x Crypto Applications

The complexity of crypto applications have long been criticized as a significant barrier to widespread adoption.

Over the last fifteen years, blockchains have required users to navigate intricate approvals, manage private keys, and understand complex UI patterns that are beyond the grasp of most internet users.

With agents, these user patterns are changing quickly.

If you view AI models as reactive infrastructure, responding to prompts based on previous training data, then you might view AI agents as proactive applications — integrating models into new architectures to accomplish narrow objectives.

Simply, AI agents leverage foundational models to autonomously think, autonomously plan, and autonomously take action.

It is important to understand that agents are unique from the “bots” we know from yesterday.

Unlike bots, AI agents can reason on demand. They can analyze their own performance, adapt their strategies, and fight through complex tasks that sometimes require hundreds — and in the future — thousands of unique steps.

In September 2024, I met with one of my portfolio founders building an AI agent protocol for navigating blockchains.

The protocol is called Wayfinder.

Using his phone, and a few short natural language prompts, I deployed a front end and token contract replicating the monetary policy of Bitcoin to the BASE blockchain, using ETH bridged from Ethereum mainnet.

The whole process took less than four prompts on Wayfinder, and a total of five minutes from start to finish.

Nascent applications like Wayfinder suggest an important trend:

AI agents will mediate all of crypto’s longstanding technological frictions.

Over the next 12 months, agents will transform blockchain’s complex structure into seamless, natural language interactions that increase protocol accessibility, safeguard users from their own mistakes, help developers deploy safer code, and dramatically lower consumer churn across a complex decentralized web of products and services.

Importantly, key management networks will extend all of these agentic utilities, allowing agents to roam seamlessly over blockchains even when humans are not at the computer. And universal namespace networks will connect every agent’s actions — across every blockchain — under one human identity.

Put simply:

Crypto agents make building or using any crypto product, across any blockchain, easier for humans.

- In decentralized finance, agents will collapse crypto’s financial friction into simple natural language intents, in integrity with each user’s risk demands.

- In crypto gaming, agentic workflows will empower personalized and ownable generative objects, more sophisticated non-player characters in cryptoeconomies, and tailored experiences for players on-chain.

- In decentralized organizations, humans will agree on overarching policy objectives and constraints at the edges — and allow the agentic middle to execute across all dimensions of business, protocol, or administrative functionality.

These navigational benefits are a radical zero to one moment for all crypto applications.

Millions of new users will onboard this way.

No category will be spared.

AI, Crypto and Ownership

Understanding the grand trend line in front of us, it’s important to take a step back and honor the lessons from our past.

For most of human history, the ability to secure and defend resources meant survival itself. Modern property ownership emerged from millions of years of this evolutionary pressure.

The concept of property rights are so integral to the human experience that they are enshrined in the US Constitution (5th Amendment). America’s founding fathers viewed property rights as fundamental to our system of governance and way of life.

Economists have also long argued that strong property rights are a cornerstone of economic growth. They are critical for allowing individuals to feel safe generating income, preserving/storing wealth, and leveraging these assets for credit and investment over time.

Research backs this up.

A study of over 100+ countries from 1990–2002 revealed that countries with stronger property rights were able to grow faster than those with weaker property rights, in part because of their effect on stimulating technology growth and improved resource allocation.

From a property rights lens, blockchains are a winning technology.

They are the strongest technological foundation for the world’s digital information — by empowering immutable record-keeping, cryptographically-secured ownership, distributed security, and programmatic enforcement of rights through smart contracts.

And as America moves into the age of digital intelligence, blockchains can also serve as the canonical environment for all AI infrastructure and applications — ensuring that America’s AI systems can benefit from the structural tailwinds of crypto’s five unique features.

Over history, America has created unprecedented opportunity for individuals and country alike — from breaking free of colonial rule, to its constitutional commitment to personal freedom, to fighting segregation — coupled with fierce market competition and an entrepreneurial focus.

Today, on the precipice of AGI, I believe America has an important opportunity to further solidify itself as a leader on these same dimensions.

Integrating America’s AI policy goals with crypto will catalyze unprecedented individual participation in open source networks, through incentivized contribution at all layers of the AI stack.

Widespread participation in our AI markets will stimulate competition, encouraging new forms of bottoms-up mobilization and broader social impact.

And finally — blockchains will provide a secure digital foundation for America’s next property rights system, facilitating a trust-minimized framework for Americans to store and preserve digital wealth across the internet.

AI, Crypto, and America.

Three of the world’s most important experiments — operating together.

So that future generations of America can dream like our parents did.

Disclosure/Disclaimer: This article does not constitute an offer to sell, or a solicitation of an offer to buy the interests of any Fund. Any such offer will be made pursuant to its offering documents. This article reflects the current opinions of the author(s), is educational in nature and should not be considered to be a recommendation to invest. Investing in digital assets, NFTs or cryptocurrency includes risk including the loss of principal. Past performance is not indicative of future results. While attempts have been made to verify the accuracy of the information provided, we cannot make any guarantees. Any forward-looking statements made are based on certain assumptions and analyses of historical trends, current conditions, and expected future developments, as well as other factors. Such statements are not guarantees of future performance and are subject to certain risks, uncertainties, and assumptions that are difficult to predict.

All sketch artwork in this essay was created by Harry Lange (1930–2008) for Stanley Kubrick’s 2001: A Space Odyssey.

Grateful to those who contributed ideas, conversation, or review during the construction of this piece — including MP, Fred Wilson, Kalos, Foobar, Stats (Punk9059), Ben Roy, Jack Butcher, David Sneider, Gmoney, and the killer crew at Collab+Currency (my business partner Steve and awesome team Amanda, Karen, Greg, Aradtski, Justin, Jerry, and Mark — with a special thank you to Zack and Ronan).

AICrypto