Funding Open-Source Generative AI With Crypto

The intersection between generative AI and Web3 has been one of the most active areas of research and development in crypto circles over the past few months. Decentralized computing, zero-knowledge AI, smaller base models, decentralized data networks, and AI-first chains are some of the latest trends aimed at enabling Web3 native rails for AI workloads.

These trends are technological innovations that aim to bridge the Web3 and AI worlds, representing a natural friction against the centralized nature of generative AI. Although building technological bridges with AI underpins the evolution of Web3, they do not represent the only integration path for these technology trends.

What if the way to integrate Web3 with AI was financial rather than purely technical? It turns out that cryptocurrency’s programmable finance and capital creation capabilities could be useful for one of the biggest challenges facing the current productive AI market.

Which challenge are we talking about? Open source generative AI is nothing but funding challenges.

Despite the recent level of innovation in decentralized generative AI, the gap with centralized AI technology is widening rather than narrowing. Many agree that blockchains represent the best technology alternative to the increasing centralized AI control of major tech platforms. However, the adoption challenges of decentralized AI platforms are huge.

Decentralized computing is a clear mainstay for decentralized AI, but has proven impractical for pre-training and fine-tuning workloads that require near-GPUs with access to datasets that often reside behind corporate firewalls. Zero-knowledge machine learning is too expensive to be practical on large baseline models and has not seen any real demand in the market. Decentralized data marketplaces need to overcome the same problems that have prevented data marketplaces from becoming big tech businesses.

As decentralized AI struggles to overcome these frictions, centralized alternatives are accelerating at a furious pace, creating a frightening gap between the two. The only trend that sustains hopes of a world where decentralized AI can thrive is the rapid development of open-source generative AI.

All decentralized AI trends rely on a healthy open-source generative AI ecosystem, but that ecosystem may not be as healthy as it seems.

Over the last few years we have witnessed an explosion of innovation in the field of open source massive generative AI as an alternative to platforms such as OpenAI/Microsoft, Google or Anthropic. Meta has become an amazing and undisputed champion of open source generative AI with the release of their Llama model. Companies like Mistral have raised billions of dollars in venture funding, enterprise platforms like Databricks or Snowflake are touting open source models, and there are a growing number of open source generative AI releases on a weekly basis.

The story continues

While momentum in open source generative AI is strong, a more detailed analysis shows a different reality. Open source generative AI faces a major funding challenge. When it comes to large off-the-shelf models, only large companies like Databricks, Snowflake, Meta, or well-funded startups like Mistral can keep up with the performance of large off-the-shelf models. While most releases from other labs like Databricks and Snowflake focus on optimized enterprise workloads, most recent open source research focuses on complementary techniques rather than new models.

The reason behind this phenomenon can be attributed to the astronomical costs of building large frontier models. Any pre-training cycle for a model with more than 20 billion parameters can cost between tens and hundreds of million dollars and involves a multi-month process with many failed attempts. These costs are outside the budget of most university laboratories. To make things even more interesting, most grants to AI university labs come from large tech companies, which in turn directly benefit from the outputs.

Making money with open source has been historically difficult, and making money with open source generative AI is also difficult at AI scale. As a result, open source generative AI is experiencing a major funding shortage that could create a serious gap with AI executives.

Cryptocurrency capital formation primitives appear to be one of the few viable alternatives to address the funding shortage in generative AI. Throughout their history, crypto tokens have been the primary vehicle for capital formation for Web3 projects through bull and bear market cycles. Could some of these principles be applied to open source generative AI? There are definitely more than one interesting option.

Gitcoin Second Order Funding

Gitcoin represents one of the most successful examples of financing open source innovation on Web3. The second-order financing mechanism pioneered by Gitcoin can be directly applied to generative AI. Bringing native generative AI capabilities to Web3 is crucial to the evolution of the field, so it’s natural to expect generative AI projects to attract the attention of the community.

Let’s say a university AI lab needs to raise $10 million for a pre-graduate course based on new architecture. Multiple DAOs and foundations can contribute to a Gitcoin donation that donors can also match, creating a more efficient funding mechanism. This mechanism is much more efficient than existing alternatives on the market.

A New Open Source Generative AI License

Financing open source projects enables mechanisms through which the value created by these projects can benefit the original funding community. An interesting idea when it comes to Web3 and open generative AI is to create a license where any commercial application using a model funded using Web3 tokens will contribute a portion of that revenue back in the form of that token. This mechanism can even be implemented with smart contracts.

Funding tools for open source AI are one of the most important challenges to be addressed in the current generative AI environment. Open source is traditionally difficult to finance, and open source generative AI is even more difficult to finance, given its expensive computational requirements.

Failure to enable appropriate funding channels to foster open source innovation in generative AI could pose a systemic risk to the entire field, as the balance will shift to fully closed commercial platforms. Crypto has established some of the most advanced and battle-tested channels for funding open source innovation. Maybe the first bridge between Web3 and generative AI will be financial, not technical.

Note: The opinions expressed in this column are those of the author and do not necessarily reflect the views of CoinDesk, Inc. or its owners and affiliates.