OpenAI vs. Intellectual Property: It’s Happening.

OpenAI is exploring 'value sharing' by taking a percentage or license from new drug discoveries made using its AI, a model which sparks controversy but is not unprecedented in medical research and could become common across industries.

• OpenAI is pursuing a 'value sharing' model, seeking a percentage of drug discoveries made using its AI.

• This strategy aims to monetize advanced compute resources by cutting out traditional investors in high-stakes research.

• Concerns exist that this IP-sharing model could expand to other industries, potentially creating a future where AI companies control all innovation.

OpenAI's CFO, Sarah Friar, recently discussed a 'value sharing' model where OpenAI would take a percentage or license from drug discoveries made using its AI, particularly with pharmaceutical companies requiring massive computing power. This proposal, initially met with disbelief and compared to existing IP agreements in academia and medical research, suggests that OpenAI could offer compute resources in exchange for equity, effectively cutting out traditional investors. While currently focused on drug discovery, there's concern this model could expand to other industries, potentially leading to a future where AI companies control all resources and retain ownership of new discoveries.

OpenAI's Value Sharing Plan

• 00:00:00 OpenAI is exploring a 'value sharing' model, specifically targeting drug discovery and medical research, where it would partner with pharmaceutical companies to discover new drugs and receive a percentage or license in return. This idea, initially reported from CFO Sarah Friar's comments at Davos, was widely misinterpreted on social media as a general claim on all user-generated IP, but is currently focused on high-compute, high-stakes research where rapid discovery is paramount.

Broader Revenue Strategy

• 00:01:25 This 'value sharing' initiative follows OpenAI's significant revenue growth and recent strategic announcements, including tripling revenue yearly, expanding an affordable $8 per month plan for market capture, and introducing advertising for free and 'go' tier users. These moves underscore a broader strategy to generate substantial revenue and solidify OpenAI's market position, with value sharing serving as a novel approach to monetize its advanced AI compute capabilities beyond direct subscription or advertising fees.

Historical Precedent & Concerns

• 00:03:13 The concept of intellectual property sharing in medical research, while seemingly wild, has precedents, such as the 2018 partnership between GSK and 23andMe for leveraging genetic insights in medicine development. Similar IP ownership structures are common in major universities like Stanford, where discoveries made even tangentially related to research while using university resources belong to the institution. However, applying this model more broadly to the private sector, where companies typically don't own employees' side projects, raises significant concerns about equitable ownership and the potential for AI companies to demand equity for access to their foundational compute resources across various industries.

Future Economic Implications

• 00:13:00 The expansion of this value-sharing model beyond drug discovery could lead to a future where major AI companies, possessing vast compute and capital, control all resources and dictate terms for innovation. In a scenario of Artificial General Intelligence (AGI) and an intelligence explosion, these companies might no longer need human partners, directly discovering and commercializing drugs, math, or software themselves. This raises concerns about a 'permanent underclass' and a freezing of socioeconomic status, as only those with compute resources would be able to participate in significant discovery and economic activity.