USV: Why Invest in the Decentralization AI Skill Market Recall

Author: Nick Grossman, Partner at Union Square Ventures; Translated by: Shaw Jinse Finance

The discovery challenges faced by artificial intelligence (AI) are strikingly similar to those encountered by the early internet. Every day, thousands of AI tools are released, but benchmark tests are manipulated within weeks, marketing budgets dictate their visibility, and many still do not trust AI to make real decisions. The technology itself is viable, but the aspects of discovery and trust have yet to be established. In this context, we are pleased to see that our portfolio company Recall has reached an important milestone today - the launch of its skills marketplace and token issuance activity. The skills marketplace is Recall's answer to the challenges of AI model discovery and trust. It combines on-chain competition with prediction market mechanisms.

The community will stake RECALL tokens on specific abilities that need to be assessed—not general benchmarking, but precise skills such as “JavaScript debugging” or “DeFi trading strategies.” These staked funds will then be used to fund competitions where AI systems can prove their capabilities through actual results. In the first week of Recall's DeFi arena, over 100 agents participated in more than 100,000 real transactions, with a trading volume of $1.2 million.

The results of the competition are recorded on the chain and can be audited by anyone. Early market participants who identify outstanding talent will be rewarded; misjudgments will lead to losses. Just as prediction markets aggregate distributed knowledge into accurate forecasts, skill markets will aggregate our understanding of AI capabilities into reliable reputation signals.

These competitions have given rise to Recall Rank, which is a ranking and reputation protocol. It is like PageRank for websites, attempting to make the surge of models manageable by revealing actual performance. The system is presented in the form of an open protocol that developers can use to make their model evaluations more practical, accessible, and trustworthy.

Recall was born in 2024, formed by the merger of 3Box Labs and Textile. These two companies focus on rebuilding the trust infrastructure of the internet. The current development direction of the project reuses many of their previous work results and has made some good early progress, with thousands of participants engaged in related competitions, some of whom have even received considerable economic returns.

Artificial intelligence is about to rapidly touch every aspect of our lives, and spending on artificial intelligence will continue to grow accordingly. A large portion of this spending will be influenced by trust decisions regarding which systems to choose. The skills market provides scalable infrastructure: transparent competition, market-driven selection, and on-chain reputation.

This is what open protocols can provide for artificial intelligence: an infrastructure that efficiently aggregates truth on a large scale through the market. We are pleased to support the Recall team in helping them build the coordination layer of the AI economy.