According to @VitalikButerin, using a single AI to allocate funding invites jailbreak exploits like gimme all the money, making naive AI governance unsafe for resource distribution, source: https://twitter.com/VitalikButerin/status/1966688933531828428. He endorses an info finance design featuring an open market where anyone can submit models, enforced by a spot-check mechanism that anyone can trigger and a human jury to evaluate results, source: https://twitter.com/VitalikButerin/status/1966688933531828428 and https://vitalik.eth.limo/general/2024/11/09/infofinance.html. He argues this plug-in marketplace is more robust because it provides real-time model diversity and creates built-in incentives for model submitters and external speculators to detect and correct issues quickly, source: https://twitter.com/VitalikButerin/status/1966688933531828428. For trading relevance, his emphasis on open markets, human-in-the-loop review, and speculator incentives highlights market-based verification as a preferred mechanism for AI funding systems that traders can monitor for adoption in governance and model marketplaces, source: https://twitter.com/VitalikButerin/status/1966688933531828428 and https://vitalik.eth.limo/general/2024/11/09/infofinance.html.