Recap: 11 Intersection Scenarios of Artificial Intelligence and Cryptocurrency
TechFlow Selected TechFlow Selected
Recap: 11 Intersection Scenarios of Artificial Intelligence and Cryptocurrency
Hope this can spark more discussions in the industry: what is feasible, what challenges remain to be solved, and how things might evolve in the future.
Navigating Web3 tides with focused insightsAuthors: Scott Duke Kominers, Sam Broner, Jay Drain, Guy Wuollet, Elizabeth Harkavy, Carra Wu, and Matt Gleason
Translated by: Aki Wu Shuo Blockchain
The economic structure of the internet is changing. As open networks gradually collapse into a "prompt bar," we must ask: Will AI lead to a more open internet, or will it trap us in a maze of new paywalls? And who will ultimately control the future internet—large centralized companies, or broad user communities?
This is where TechFlow comes in. We've discussed the intersection of AI and crypto many times before, but in short, blockchain offers a way to redesign internet services and network architecture to build decentralized, credibly neutral, and user-"ownable" systems. By reshaping the economic incentives underpinning today’s systems, blockchain counterbalances the growing centralization within AI, paving the way for a more open and resilient internet.
The idea that “crypto can help build better AI systems, and vice versa” isn’t new—but it has long lacked clear definition. Certain crossover areas (such as verifying "human identity" amid an explosion of low-cost AI systems) have already attracted significant developer and user interest. But other use cases may take years or even decades to materialize. This article outlines 11 convergence points between AI and crypto, aiming to spark broader industry discussion: which are feasible, what challenges remain, and how these might evolve.
These scenarios are all grounded in technologies currently under development—from enabling massive micropayments to ensuring humans retain ownership in their future relationship with AI.
1. Persistent Data and Context in AI Interactions
Scott Duke Kominers: Generative AI relies fundamentally on data, but in many applications, "context"—the state and background information tied to interactions—is often just as important, if not more so, than the data itself.
Ideally, agents, LLM interfaces, or other AI applications should remember vast amounts of personalized information—like your ongoing projects, communication habits, or preferred programming languages. In reality, users often have to rebuild this context repeatedly—not only when starting a new session within the same app (e.g., opening a new ChatGPT or Claude window), but especially when switching between different AI systems.
Currently, context from one generative AI application rarely transfers to another.
With blockchain, AI systems can store key contextual elements as persistent digital assets, loadable at the start of a session and seamlessly transferable across AI platforms. Moreover, because "forward compatibility" and "interoperability guarantees" are core features of blockchain protocols, blockchain may be the only technological path capable of systematically solving this problem.
A straightforward application lies in AI-driven gaming and media, where user preferences (like difficulty settings or control layouts) could persist across games and environments. But higher-value use cases involve knowledge work—where AI needs to understand a user's knowledge base, learning style, and capabilities—and specialized domains like programming assistance. While some enterprises have built custom AI tools with "global context," even these contexts fail to migrate effectively across different AI systems used within the same organization.
Organizations are only beginning to grasp this issue. The closest current solution is custom bots with fixed, persistent context. However, off-chain context portability between users is already emerging; for example, on Poe, users can rent out their self-created bots to others.
If such activities move on-chain, the AI systems we interact with could share a context layer composed of key elements from our entire digital behavior. AI could instantly understand our preferences, enabling finer personalization and optimized experiences. Conversely, mechanisms like on-chain IP registration could enable new, more robust market dynamics around prompts and information modules—if AI can reference on-chain persistent context, users could monetize their expertise through licensing while retaining full data autonomy.
As context-sharing capabilities grow, they will unlock countless unforeseen use cases and possibilities.
2. Universal Identity for Agents
Sam Broner: Identity—the standardized record of “who or what something is”—is the foundational infrastructure supporting today’s digital discovery, aggregation, and payment systems. But because platforms keep this “plumbing” locked internally, users typically experience identity only through finished product interfaces. For example, Amazon assigns identifiers (like ASIN or FNSKU) to products, aggregates them in unified views, and facilitates discovery and payment. Facebook works similarly: user identity determines news feed content and underpins discovery across various in-app content, including Marketplace listings, organic posts, and ads.
With the rapid evolution of AI Agents, this is about to change. More companies are deploying agents for customer service, logistics, and payments. Their platforms will no longer be traditional “single-interface apps,” but distributed across multiple channels, accumulating deep context and executing increasingly complex tasks on behalf of users. But if an agent’s identity is tied to a single platform or marketplace, it becomes unusable in critical environments like email threads, Slack channels, or inside other products.
Agents need a unified, portable “digital passport.” Without it, we can’t determine how to pay an agent, verify its version, query its capabilities, identify who it represents, or track its reputation across applications and platforms. An agent’s identity system must function simultaneously as a wallet, API registry, changelog, and social reputation proof—so any interface (email, Slack, or another agent) can parse and communicate with it consistently.
Without this shared “identity primitive,” every integration requires rebuilding the plumbing from scratch; discovery remains ad hoc; and users lose critical context when switching between channels and platforms.
We now have the chance to design agent infrastructure from first principles. So the question is: How do we build an identity layer richer than DNS records and credibly neutral? Instead of recreating monolithic platforms that bundle identity, discovery, aggregation, and payment, we should let agents receive payments autonomously, publicly list their capabilities, and exist across ecosystems without fear of platform lock-in.
This is precisely where TechFlow and AI converge—blockchain networks offer permissionless composability, empowering developers to create more powerful agents and better user experiences.
Vertically integrated solutions like Facebook or Amazon currently offer superior UX because building great products requires top-down coordination of all components. But the cost of this convenience is rising, especially as software costs for building, aggregating, promoting, commercializing, and distributing agents fall, and agent reach expands.
Achieving vertical-integration-level UX still demands immense effort. But once a credibly neutral agent identity layer exists, entrepreneurs truly own their passports—unlocking widespread experimentation and innovation in distribution and interaction design.
3. Future-Proof Proof of Personhood (PoP)
Jay Drain Jr. and Scott Duke Kominers: With the rise of AI—whether bots and agents interacting across web interfaces, or deepfakes and social media manipulation—it’s becoming harder to tell whether we’re interacting with real humans online. This erosion of trust isn’t a future concern—it’s happening now. From bot armies on X to automated profiles on dating apps, real and fake are blurring. In this environment, Proof of Personhood (PoP) is becoming essential internet infrastructure.
One way to verify “you are human” is through digital IDs, such as those issued by centralized authorities like TSA. Digital IDs encompass all information users can leverage to prove identity—usernames, PINs, passwords, and third-party attestations (e.g., nationality, credibility, creditworthiness). The value of decentralization here is clear: when identity data resides in centralized systems, issuers can revoke access, charge fees, or assist surveillance. Decentralization flips this—users, not platform gatekeepers, control their identities, making them more secure and censorship-resistant.
Unlike traditional identity systems, decentralized PoP mechanisms (like Worldcoin’s World ID) let users manage their own data and verify their humanity in privacy-preserving, credibly neutral ways. Like a driver’s license—usable anywhere regardless of where it was issued—decentralized PoP can serve as a universal foundational module, reusable on any platform, including those not yet built. In other words, blockchain-based PoP offers “forward compatibility” because it provides:
Portability: The protocol is an open standard, integrable by any platform. Decentralized PoP can be managed via public infrastructure and fully controlled by users—making PoP inherently portable, compatible with any present or future platform.
Permissionless Accessibility: Platforms can choose whether to support a PoP identity without going through centralized API approvals that may impose discriminatory restrictions on different use cases.
The core challenge in this space is adoption. No large-scale PoP application exists yet in the real world. But we expect PoP adoption to accelerate significantly once user numbers reach critical mass, early partners emerge, and “killer apps” drive demand. Every app adopting a digital ID standard increases its value to users, prompting more users to acquire it, which in turn makes integrating that ID standard more attractive for other apps to verify “personhood.” (Moreover, because on-chain IDs are designed for interoperability, this network effect can spread rapidly.)
We’re already seeing mainstream consumer apps in gaming, dating, and social media announce partnerships with World ID to ensure users are interacting with real humans during gameplay, chats, or transactions—even specific individuals they intend to engage with. New identity protocols have also emerged this year, such as Solana Attestation Service (SAS). While SAS itself isn’t a PoP issuer, it allows users to privately link off-chain data (like KYC results or investor accreditation) to a Solana wallet, forming a decentralized identity. These signs suggest the tipping point for decentralized PoP may be near.
Proof of Personhood is about far more than “stopping bots.” It aims to draw a clear boundary between AI agents and human networks, enabling users and apps to distinguish between human and machine interactions—laying the foundation for higher-quality, safer, and more authentic digital experiences.
4. Decentralized Physical Infrastructure for AI (DePIN)
Guy Wuollet: AI is a digital service, but its growth is increasingly constrained by physical infrastructure. Decentralized Physical Infrastructure Networks (DePIN)—a new paradigm for building and operating real-world systems—could democratize the compute infrastructure powering AI innovation, making it cheaper, more resilient, and censorship-resistant.
Why? The two main bottlenecks in AI development are energy and chip access. Decentralized energy systems can supply more abundant power, while developers are already using DePIN to aggregate idle chips from gaming PCs, data centers, and other sources. These computing devices can collectively form a permissionless compute marketplace, leveling the playing field for building new AI products.
Other applications include distributed training and fine-tuning of large language models (LLMs), and building decentralized inference networks. Decentralized training and inference can drastically reduce costs by utilizing otherwise idle compute resources. Additionally, such architectures are naturally censorship-resistant, ensuring developers aren’t “deplatformed” or restricted due to reliance on hyperscalers (centralized cloud providers offering massive scalable compute).
The concentration of AI models in a few companies has long been a concern. Decentralized networks can help build AI systems that are cheaper, more censorship-resistant, and more scalable.
5. Infrastructure and Security for Interactions Between AI Agents, Service Providers, and Users
Scott Duke Kominers: As AI tools grow more capable at handling complex tasks and multi-step interaction chains, they will increasingly need to collaborate independently with other AIs without direct human oversight.
For instance, an AI agent may need to request specific data required for a computation, or invoke another specialized agent to perform a task—like asking a statistical analysis agent to build and run model simulations, or calling on an image generation agent to assist in creating marketing materials. AI agents will also deliver immense value in end-to-end transaction execution, such as completing a purchase flow entirely on a user’s behalf: finding and booking flights based on preferences, or automatically discovering and buying new books matching user tastes.
Today, there is no “generalized agent-to-agent marketplace.” Cross-agent requests are typically limited to explicit API calls or confined within closed AI agent ecosystems as internal functions.
More broadly, most AI agents currently operate in isolated ecosystems: APIs are relatively closed, and there’s a lack of unified architectural standards. Blockchain technology can help establish open standards—critical for short-term adoption and, in the long term, forward compatibility: as new agents emerge, they can all plug into the same underlying network. Because blockchains are interoperable, open-source, decentralized, and generally easier to upgrade, they are better equipped to adapt to the evolving landscape of AI innovation.
Several companies are already building on-chain infrastructure for agent interactions. Halliday, for example, recently launched a protocol providing standardized cross-chain architecture for AI workflows and interactions, with built-in safeguards to ensure agents don’t act beyond user intent. Meanwhile, projects like Catena, Skyfire, and Nevermind use blockchain to enable automatic settlements between agents, allowing AI-to-AI payments without human intervention. Similar systems are emerging, and Coinbase has begun offering infrastructure support for such development.
6. Synchronizing AI “Vibe Coding” Applications
Sam Broner and Scott Duke Kominers: The generative AI revolution has made software development easier than ever. Coding speed has increased exponentially, and crucially, coding can now be done directly through natural language—enabling even inexperienced developers to replicate existing programs or build new applications from scratch.
However, AI-assisted coding brings significant “entropy” both within and across programs. “Vibe coding” abstracts away the complex dependencies behind software—but precisely because of this, when underlying source libraries or inputs change, programs may face functional and security risks. Additionally, when people use AI to create highly personalized apps and workflows, integration with others’ systems becomes harder. In fact, even two vibe-coded programs performing nearly identical tasks may have completely different logic and output structures.
Traditionally, file formats, operating systems, and later shared software and API integrations ensured consistency and compatibility. But in a world where software evolves, morphs, and branches in real time, standardization layers must be widely accessible, continuously upgradable, and trustworthy. Moreover, AI alone cannot solve incentive problems—specifically, how to incentivize developers to build and maintain connections between systems.
Blockchain can address both issues by providing protocolized synchronization layers embedded within user-customized software builds, dynamically updated as environments change to ensure cross-system compatibility.
In the past, large enterprises might pay millions to system integrators like Deloitte to customize a Salesforce instance. Today, an engineer might build a custom “sales data viewer” over a weekend. But as the number of customized software solutions grows, developers will need help ensuring these apps stay synchronized and usable.
This resembles today’s open-source library development, but with a key difference: synchronization layers update continuously rather than through periodic releases—and come with built-in incentives. Both can be more easily achieved via crypto. Like other blockchain-based protocols, shared ownership of synchronization layers incentivizes continuous investment in improvements. Developers, users (and their AI agents), and other participants can earn rewards for introducing, using, or iterating new features and integrations.
Conversely, shared ownership gives all users a stake in the protocol’s overall success, creating disincentives for harmful deviations. Just as Microsoft wouldn’t casually break .docx standards—due to widespread negative impacts on users and brand—co-owners of synchronization layers would avoid introducing clumsy or malicious code due to self-interest.
Like previous software standardization architectures, this approach holds strong network effects. As AI-generated software undergoes a “Cambrian explosion,” the number of diverse, heterogeneous systems needing to communicate will grow exponentially. In short: vibe coding can’t stay in sync by vibes alone—crypto is the answer.
7. Micropayment Systems Supporting Revenue Sharing
Liz Harkavy: AI agents and tools like ChatGPT, Claude, and Copilot offer easier ways to access information in the digital world. But for better or worse, they’re also disrupting the open internet’s economic structure. This trend is already visible—educational platforms are seeing significant traffic declines as students adopt AI tools, and several U.S. media outlets are suing OpenAI over copyright infringement. If incentive systems aren’t rebalanced, we may see the internet become further enclosed, with more paywalls and fewer content creators.
Policy solutions exist, but while legal processes unfold, technical alternatives are emerging. The most promising (and technically challenging) is embedding “revenue sharing” into the internet’s foundational architecture. When an AI-driven action leads to a sale, content creators whose information contributed to that decision should receive a cut. Affiliate marketing ecosystems already perform similar attribution and revenue sharing; more advanced systems could automatically track and reward all contributors along an information chain. Blockchain is clearly well-suited to tracking “provenance chains” of information.
But realizing this system requires new infrastructure—especially: micropayment systems capable of processing extremely small amounts across multiple sources; attribution protocols that fairly assess contribution value; and governance models ensuring transparency and fairness.
Many existing blockchain tools show promise—various rollups, L2 networks, AI-native financial institutions like Catena Labs, and financial infrastructure protocols like 0xSplits—all enabling near-zero-cost transactions and finer-grained payment splits.
Blockchain enables sophisticated, agent-driven payment systems through multiple mechanisms:
Nanopayments: Can be split among multiple data providers, allowing a single user interaction to automatically trigger micro-payments to all contributing sources via smart contracts.
Smart Contracts: Can automatically execute enforceable “after-the-fact payments” post-transaction, providing transparent, traceable compensation to content sources influencing purchase decisions.
Programmable Payment Splits: Enable revenue distribution enforced by code, not centralized entities, establishing trustless financial relationships between automated agents.
As these technologies mature, they will form a new media economy model, capturing the full value creation chain—from creators to platforms to users.
8. Blockchain as a Registration System for Intellectual Property and Provenance
Scott Duke Kominers: The rise of generative AI makes efficient, programmable intellectual property (IP) registration and tracking urgent—not only to ensure accurate provenance, but also to support new business models around IP access, sharing, and derivative creation. Existing IP frameworks rely on costly intermediaries and reactive enforcement—clearly inadequate in an era where AI instantly consumes content and generates variants with one click.
We need open, public registries that provide creators with clear ownership proofs, low barriers to entry, and high efficiency—while also being directly interactive with AI and other web applications. Blockchain is ideal for this role: it allows creators to register IP without intermediaries, provides immutable provenance records, and enables third-party apps to easily identify, license, and interact with these IP assets.
Of course, skepticism remains about whether “technology can truly protect IP.” After all, previous internet eras—and even the current AI revolution—often correlate with weakened IP protection. One reason is that many existing IP business models emphasize “excluding derivatives” rather than incentivizing or monetizing them. Programmable IP infrastructure can not only help creators, franchises, and brands assert IP ownership in digital spaces, but also spawn new business models centered on “sharing IP for generative AI and digital applications.” In a sense, it turns one of generative AI’s threats to creative work into an opportunity.
In NFT’s early days, we saw creators experiment with new models—like building brand network effects via CC0 on Ethereum to capture value. Recently, infrastructure providers have started building standardized, composable IP registration and licensing protocols, even launching dedicated blockchains (like Story Protocol). Some artists now use protocols like Alias, Neura, and Titles to license their styles and works for creative remixing. Meanwhile, Incention’s sci-fi series Emergence invites fans to co-create universe and character lore, with each contribution recorded via on-chain registration on Story.
9. Web Crawlers That Compensate Content Creators
Carra Wu: The most product-market-fit AI agents today aren’t those for programming or entertainment—they’re web crawlers that autonomously browse the internet, collect data, and decide which links to follow.
Estimates suggest nearly half of today’s internet traffic already comes from non-human sources. Bots often ignore robots.txt files—the standard meant to inform automated crawlers whether sites permit access—but in practice, these files have little enforceability. These bots use scraped data to strengthen the core moats of the world’s largest tech companies. Worse, websites bear the cost—spending bandwidth and CPU resources serving endless anonymous crawlers. In response, companies like Cloudflare and other CDNs offer blocking services. Together, this forms a patchwork system that shouldn’t exist.
We’ve previously noted that the internet’s original social contract—content creators make content, platforms distribute it—is unraveling. This is evident in the data: over the past twelve months, website operators have increasingly blocked AI crawlers. In July 2024, only about 9% of the global top 10,000 websites blocked AI crawlers; today, that figure is 37%. As more site owners gain technical sophistication and user frustration grows, this number will keep rising.
So what if, instead of paying CDNs to blanket-block suspected bots, we tried a middle path? What if AI crawlers stopped freeloading and paid for data access? Here, blockchain can help: in this vision, each crawler agent holds crypto assets and negotiates on-chain with website “gatekeeper agents” or paywall protocols via the x402 protocol. (Of course, the challenge lies in robots.txt—the “Robots Exclusion Standard”—deeply entrenched in internet operations since the 1990s. Changing this will require massive collaboration or CDN support from firms like Cloudflare.)
Meanwhile, human users could prove they’re real via World ID (see above) and gain free access. Thus, content creators and site operators could be compensated the moment AI scrapes their data, while human users continue enjoying a freely flowing internet.
10. Privacy-Preserving Ads That Are Precise but Not Creepy
Matt Gleason: AI is already shaping how we shop online. But what if the ads we see daily were actually useful? People dislike ads for many reasons: irrelevant ads are pure noise; and not all “personalization” is good. Highly targeted ads driven by vast consumer data feel invasive; others try monetization through “forced ad views” (like unskippable ads in streaming or game levels).
Crypto can help fix these issues, offering a chance to reimagine advertising. When AI agents and blockchain combine, they can tailor ads based on user-defined preferences—making ads relevant without being unnervingly “creepy.” More importantly, user data stays private globally, and users willing to share data or engage with ads can be compensated.
Realizing this model requires several technical foundations:
Low-Cost Digital Payments: To compensate users for ad interactions (views, clicks, conversions), businesses need to send numerous tiny payments. Scaling this requires fast, high-throughput systems with near-zero fees.
Privacy-Preserving Data Verification: AI agents need to verify whether consumers meet certain demographic criteria. Zero-knowledge proofs (ZKP) can perform such verification without revealing private details.
New Incentive Models: If the internet adopts micro-payment-based monetization (e.g., <$0.05 per interaction), users could opt in to view ads for compensation—transforming today’s “data extraction model” into a “user participation model.”
For decades, efforts have aimed to make ads more “relevant”—online and offline. Reimagining advertising through the lens of crypto and AI can finally make ads useful, controllable, and optional. For builders and advertisers, this means more sustainable, consistent incentives; for users, richer ways to discover information and explore the digital world.
In the end, this won’t just make ad inventory more valuable—it could disrupt today’s entrenched, extractive ad economy, replacing it with a more human-centric system where users aren’t the “product sold,” but true participants.
11. User-Owned and Controlled AI Companions
Guy Wuollet: Many people now spend more time on devices than in face-to-face interactions, and that screen time is increasingly spent engaging with AI models or AI-curated content. These models already offer a form of “companionship”—for entertainment, information, niche interests, or as educational tools for children. It’s easy to imagine that in the near future, AI companions for education, healthcare, legal advice, or even emotional support will become primary modes of human interaction.
Future AI companions will offer infinite patience and deep personalization tailored to individuals and their contexts. They won’t just be assistants or “robotic servants,” but potentially deeply valued relational entities. Hence the question: Who owns and controls these relationships—users, or companies and intermediaries? If you’ve been concerned about social media content curation and moderation over the past decade, this issue will soon become exponentially more complex and personal.
The argument that “anti-censorship hosting platforms like blockchain may be the best path to uncensorable, user-controlled AI” has been well-made. While users could run local models or buy GPUs, for most people this is either too expensive or too technically demanding.
Though widespread AI companions are still some distance away, the technology is maturing fast: text chat AI is already remarkably natural; visual avatars are improving; blockchain performance is advancing. To make “uncensorable AI companions” truly user-friendly, we’ll need better crypto UX. Fortunately, wallets like Phantom have simplified blockchain interactions, while embedded wallets, Passkeys, and account abstraction now let users self-custody without managing seed phrases. Meanwhile, high-throughput, trustless computing systems based on optimistic and ZK coprocessors will enable meaningful, sustainable long-term relationships with digital companions.
In the near future, public discourse will shift from “when will realistic digital companions and avatars arrive?” to “who will control them, and how?”
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
a16z
