In the mixed landscape of Crypto x AI projects, how to identify genuine scenarios and false demands?
Cryptography is seen as a force against the centralization of AI.Original Title: AI <> Crypto Projects That Aren't Complete Bullshit
Original Author: 563, former Bankless researcher
Original Compilation: Deep Tide TechFlow
Navigating the intersection of crypto and artificial intelligence.
In the search for new alpha information, we inevitably encounter some junk. When a project can quickly raise five to six figures with just a half-clear introduction and some decent branding, speculators latch onto every new narrative. And as traditional finance joins the AI craze, the "crypto AI" narrative exacerbates this issue.
Most of these projects have the following problems:
Most crypto projects do not need AI
Most AI projects do not need cryptocurrency
Not every decentralized exchange (DEX) needs a built-in AI assistant, and not every chatbot needs a companion token to drive its adoption curve. This rigid coupling of AI and crypto nearly made me collapse when I first delved into this narrative.
What's the bad news? Continuing down the current path will further centralize this technology, ultimately leading to failure, while a plethora of fake "AI x Crypto" projects will hinder our ability to turn the tide.
What's the good news? There is light at the end of the tunnel. Sometimes, AI can indeed benefit from crypto economics. Similarly, in some cryptocurrency use cases, AI can solve real problems.
In today's article, we will explore these key intersections. The overlap of these niche innovative ideas forms a whole that is greater than the sum of its parts.
A High-Level View of the AI Stack
Here’s my perspective on the different verticals within the "crypto AI" ecosystem (if you want to dive deeper, check out Tommy's article). Note that this is a very simplified view, but I hope it helps us lay the groundwork.
At a high level, here's how it works together:
Massively collecting data.
Processing that data to enable machines to understand how to ingest and apply it.
Training models on that data to create a general model.
This model can then be fine-tuned to handle specific use cases.
Finally, these models are deployed and hosted so that applications can query them for useful implementations.
All of this requires substantial computational resources, which can run locally or be sourced from the cloud.
Let’s explore each of these areas, with a particular focus on how different crypto economic designs can actually improve standard workflows.
Crypto Gives Open Source a Fighting Chance
The debate between "closed source" and "open source" development methods can be traced back to the Windows-Linux debate and Eric Raymond's famous "The Cathedral and the Bazaar" theory. While Linux is widely used among enthusiasts today, about 90% of users choose Windows. Why? Because of incentives.
At least from the outside, open-source development has many advantages. It allows the maximum number of people to participate in and contribute to the development process. But in this headless structure, there is no unified directive. CEOs do not proactively encourage as many people as possible to use their products to maximize their bottom line. In open-source development, projects can evolve into a "chimera," splitting off in different directions at every intersection of design philosophy.
What’s the best way to adjust incentives? Build a system that rewards behaviors that facilitate goal achievement. In other words, put money in the hands of those who can bring us closer to our goals. With cryptocurrency, this can be hard-coded into law.
We will look at some projects that are doing just that.
Decentralized Physical Infrastructure Networks (DePINs)
"Oh come on, not this again?" Yes, I know the DePIN narrative has been told almost as much as AI itself, but bear with me for a moment. I am willing to believe that DePINs are a genuinely promising crypto use case that can change the world. Think about it.
What is crypto really good at? Removing intermediaries and incentivizing activities.
The original vision of Bitcoin was peer-to-peer currency, designed to exclude banks. Similarly, modern DePINs aim to exclude centralized powers and introduce provably fair market dynamics. As we will see, this architecture is ideal for crowdsourcing AI-related networks.
DePINs use early token issuance to increase the supply side (providers), hoping to attract sustainable consumer demand. This aims to solve the cold start problem of new markets.
This means that early hardware/software ("nodes") providers earn a lot of tokens and a little cash. As users leverage the cash flow brought by these nodes (in our case, machine learning builders), this begins to offset the decreasing token issuance over time until a fully self-sustaining ecosystem is established (which may take years). Early adopters like Helium and Hivemapper demonstrate the effectiveness of this design.
Data Networks, The Case of Grass
GPT-3 is said to have been trained on 45TB of pure text data, equivalent to about 90 million novels (but it still can't draw a circle). The data requirements for GPT-4 and GPT-5 are even greater than what exists on the surface web, making it an understatement to call artificial intelligence "data-hungry" this century.
If you are not a top player (OpenAI, Microsoft, Google, Facebook), acquiring this data is very difficult. The common strategy for most is web scraping, which is fine until you try to scale it up. If you try to scrape a large number of websites using an Amazon Web Services (AWS) instance, you will quickly hit rate limits. This is where Grass comes into play.
Grass connects over two million devices, organizing them to scrape websites from users' IP addresses, collecting, structuring, and selling that data to AI companies in desperate need of it. In return, users participating in the Grass network can earn stable income from the AI companies using their data.
Of course, there is currently no token, but a future $GRASS token could incentivize users to download their browser extension (or mobile app). They have already attracted a large number of users through an extremely successful referral campaign.
GPU Networks, The Case of io.net
Perhaps even more important than data is computational power. Did you know that in 2020 and 2021, China spent more on GPUs than on oil? This is just crazy, but it's only the beginning. Goodbye oil coins, hello compute coins.
There are many GPU DePINs on the market, and they generally work as follows:
Machine learning engineers/companies in urgent need of computation.
On the other hand, data centers, idle mining rigs, and hobbyists with idle GPUs / CPUs.
Despite the vast global supply, there is a lack of coordination. It’s not easy to contact 10 different data centers to get them to bid for your usage. A centralized solution would create a rent-seeking intermediary whose incentive is to extract maximum value from each party, but crypto can help.
Crypto is very good at creating market layers that can efficiently connect buyers and sellers. A code snippet does not need to be accountable to shareholders' financial interests.
io.net stands out because it introduces some cool new technologies that are crucial for AI training—namely, their cluster stack. Traditional clusters involve physically connecting a bunch of GPUs in the same data center to work together for model training. But what if your hardware is distributed globally? io.net has developed cluster middleware in collaboration with Ray (used to create ChatGPT) that can connect GPUs located in different places.
Moreover, while the AWS registration process can take days, clusters on io.net can be launched without permission in 90 seconds. For these reasons, I can see io.net becoming the hub for all other GPU DePINs, all of which can plug into their "IO engine," unlocking built-in clusters and a smooth onboarding experience. All of this is only possible with the help of crypto technology.
You will notice that most ambitious decentralized AI projects (such as Bittensor, Morpheus, Gensyn, Ritual, Sahara) have clear "computational" needs—this is precisely where GPU DePINs should fit in, as decentralized AI requires permissionless computation.
Application of Incentive Structures
Back to the insights from Bitcoin. Why do miners continuously compute hashes so quickly? Because that’s how they get rewarded—Satoshi proposed this architecture because it optimizes for security. What’s the lesson? The incentive structures built into these protocols determine the final products they produce.
Bitcoin miners and Ethereum stakers are participants who absorb all their native tokens because that’s what the protocol wants to incentivize—participants become miners and stakers.
In an organization, this may come from the CEO, who defines the "vision" or "mission statement." But humans are prone to error and can lead a company off track. On the other hand, computer code can maintain focus better than the roughest wage slave. Let’s look at a few decentralized projects where their built-in token effects keep participants focused on noble goals.
AI Building Networks, Exploring Bittensor
What if we let Bitcoin miners build AI instead of solving useless math problems? That’s how you get Bittensor.
Bittensor aims to create several experimental ecosystems for experimentation, with the goal of producing "commoditized intelligence" within each ecosystem. This means one ecosystem (called a subnet, or "SN") might focus on developing language models, another on financial models, and more on voice synthesis, AI detection, or image generation (see currently active projects).
For the Bittensor network, what you want to do doesn’t matter. As long as you can prove your project is worth funding, the incentives will flow. This is the goal of the subnet owner, who registers the subnet and adjusts the rules of the game.
The participants in this "game" are called miners. These are ML/AI engineers and teams building models. They are locked in a continuously reviewed "thunderdome," competing against each other for the most rewards.
Validators are another facet, responsible for reviewing and scoring the miners' work accordingly. If a validator is found colluding with a miner, they will be expelled.
Remember the incentives:
Miners earn more when they outperform other miners within the subnet—this drives AI development.
Validators earn more when they accurately identify high-performing and low-performing miners—this maintains the fairness of the subnet.
Subnet owners earn more when the AI models produced by their subnet are more useful than those from other subnets—this drives subnet owners to optimize their "game."
You can think of Bittensor as a perpetual reward machine for AI development. Emerging machine learning engineers can try to build something, pitch it to VCs, and try to raise some funds. Or they can join one of the Bittensor subnets as miners, showcase their skills, and earn a lot of TAO. Which is easier?
Some top teams are building on the network:
Nous Research is the king of open source. Their subnet has disrupted traditional methods in fine-tuning open-source LLMs. They make leaderboards unmanipulatable by continuously testing models with synthetic data flows (unlike traditional benchmarks like HuggingFace).
Taoshi's proprietary training network is essentially an open-source quantitative trading company. They ask ML contributors to build trading algorithms that predict asset price movements. Their API provides quant-level trading signals for retail and institutional users and is rapidly moving towards significant profitability.
Cortex.t, developed by the Corcel team, has two purposes. First, they incentivize miners to provide API access to top models (like GPT-4 and Claude-3) to ensure ongoing availability for developers. They also provide synthetic data generation, which is very useful for model training and benchmarking (which is also why Nous uses it). Check out their tools—chat and search.
Unsurprisingly, Bittensor reaffirms the power of incentive structures, all made possible by crypto economics.
Intelligent Agents, Exploring Morpheus
Now, let’s look at two aspects of Morpheus:
Crypto economic structures are building AI (crypto helps AI)
AI-enabled applications are enabling new use cases in crypto (AI helps crypto)
"Intelligent agents" are simply AI models trained on smart contracts. They understand the inner workings of all top DeFi protocols, know where to find yields, where to bridge, and how to discover suspicious contracts. They are the future "auto routers," and in my view, they will be how everyone interacts with blockchains in 5-10 years. In fact, once we reach that point, you might not even know you are using crypto. You will simply tell a chatbot that you want to move some savings into another investment, and everything will happen in the background.
Morpheus embodies the message of "incentivize them, and they will come." Their goal is to have a platform where intelligent agents can proliferate and thrive, with each agent building on the success of the previous one, in an ecosystem that minimizes externalities.
The token inflation structure highlights four main contributors to the protocol:
Code------agent builders.
Community------building front-end applications and tools to attract new users to the ecosystem.
Computation------providing the computational power to run agents.
Capital------providing their yields to drive Morpheus's economic machine.
Each of these categories receives an equal share of the $MOR inflation rewards (with a small portion also saved as an emergency fund), forcing them to:
Build the best agents------when their agents are consistently used, creators are rewarded. Unlike freely providing OpenAI plugins, this method pays builders immediately.
Build the best front-end/tools------when their creations are consistently used, creators are rewarded.
Provide stable computational power------providers are rewarded when they lend out computational power.
Provide liquidity for the project------by maintaining liquidity for the project, they earn their share of MOR.
While there are many other AI/intelligent agent projects, Morpheus's token economic structure is particularly clear and effective in designing incentive mechanisms.
These intelligent agents are the ultimate example of AI truly eliminating barriers to crypto applications. The user experience of dApps is notoriously poor (although there have been many improvements over the past few years), and the rise of LLMs has ignited the passion of everyone wanting to be a founder in Web2 and Web3. Despite the plethora of profit-driven projects, excellent projects like Morpheus and Wayfinder (see demo below) demonstrate how much simpler on-chain transactions will become in the future.
Putting it all together, the interactions between these systems might look something like this. Note that this is an extremely simplified view.
How to Tell if a Project is Completely Useless
Remember our two broad categories of "crypto x AI":
Crypto helps AI
AI helps crypto
In this article, we primarily explored the first category. As we have seen, a well-designed token system can lay the groundwork for the success of the entire ecosystem.
First Category - Crypto Helps AI
DePIN architecture can help kickstart markets, and creative token incentive structures can coordinate open-source projects toward goals that were once difficult to achieve. Yes, there are several other legitimate intersections that I haven’t covered due to space constraints:
Decentralized storage
Trusted execution environments (TEE)
Real-time data acquisition (RAG)
Zero-knowledge x machine learning for inference/origin verification
When deciding whether a new project is truly valuable, ask yourself:
If it is a derivative of another mature project, is its difference enough to stand out?
Is it just a repackaged version of open-source software?
Does this project genuinely benefit from crypto technology, or is crypto technology being force-fitted into it?
Do we really need 100 crypto projects like HuggingFace (a popular open-source machine learning platform)?
Second Category - AI Helps Crypto
In this category, I personally see more fake projects, but there are indeed some cool use cases. For example, AI models can eliminate barriers in the crypto user experience, especially intelligent agents. Here are some interesting categories worth watching in AI-supported crypto applications:
Enhanced intent systems------automating cross-chain operations
Wallet infrastructure
Real-time alert infrastructure for users and applications
If it’s just a "chatbot with a token," to me, that’s a garbage project. Please stop hyping these projects to maintain my sanity. Also:
Adding AI won’t magically give your failing application/chain/tool product-market fit
No one will play a bad game just because it has an AI character
Slapping an "AI" label on your project doesn’t make it interesting
Where Do We Go from Here
Despite the noise, some serious teams are working hard to realize the vision of "decentralized AI," which is worth striving for.
In addition to projects incentivizing open-source model development, decentralized data networks open new doors for emerging AI developers. When most of OpenAI's competitors struggle to reach large-scale deals with Reddit, Tumblr, or WordPress, distributed scraping can balance this gap.
A company's computational power may never exceed the total computational power owned by other companies in the world, and with decentralized GPU networks, this means anyone else has the capability to compete with top companies. All you need is a crypto wallet.
Today we stand at a crossroads. If we focus on those truly valuable "crypto x AI" projects, we have the ability to decentralize the entire AI stack.
The vision of cryptocurrency is to create a hard currency that no one can interfere with through the power of cryptography. Just as this emerging technology begins to gain traction, a more formidable challenger has emerged.
Ideally, centralized AI will not only control your finances but will also impose biases on every piece of data we encounter in our daily lives. It will enrich a select few tech leaders in a self-perpetuating cycle of data collection, fine-tuning, and model injection.
It will know you better than you know yourself. It knows which buttons to press to make you laugh more, get angrier, and consume more. Despite appearances, it is not accountable to you.
Initially, crypto technology was seen as a force against AI centralization. Crypto has the ability to coordinate decentralized individuals to work together toward a common goal. However, this ability is now facing a more powerful enemy than central banks: centralized AI. This time, time is of the essence, and we need to act swiftly to resist the trend of AI centralization.
