Microsoft and OpenAI have a financial definition of AGI: Report

Microsoft and OpenAI have a unique definition of artificial general intelligence (AGI) tied to financial performance, as reported by The Information. According to an agreement between the two companies, OpenAI will have achieved AGI when its AI systems generate at least $100 billion in profits. This definition diverges from the traditional technical and philosophical understanding of AGI. Currently, OpenAI is projected to incur significant losses and does not expect to be profitable until 2029. This financial-based definition of AGI is crucial because Microsoft will lose access to OpenAI's technology once AGI is declared.

The agreement implies that Microsoft could continue to access OpenAI's models for many years, as OpenAI is far from meeting the profit threshold. There is speculation that OpenAI might declare AGI prematurely to limit Microsoft's access, but the current agreement suggests otherwise. Recently, discussions have arisen about whether OpenAI's o3 model represents progress toward AGI. Despite its improved performance, the model's high compute costs challenge the profit-driven AGI definition set by OpenAI and Microsoft.

Key takeaways

Microsoft and OpenAI have defined AGI based on the startup's ability to generate $100 billion in profits.
OpenAI is reportedly set to lose billions this year and does not expect to turn a profit until 2029.
Microsoft loses access to OpenAI's technology once AGI is achieved, according to their agreement.
There is debate over whether OpenAI's o3 model is a significant step toward AGI, given its high compute costs.

Microsoft and OpenAI have a financial definition of AGI: Report | TechCrunch

Key takeaways

Discussion (0)