In the next week, OpenAI will create a marketplace for GPTs, bespoke apps based on its text-generating AI models (e.g., GPT-4).
OpenAI informed Eltrys in an email that developers constructing GPTs must examine the company’s new usage restrictions and GPT brand standards to guarantee compliance before placing them in the GPT Store. They must also validate their user profile and post GPTs as “public.”
Last year, during OpenAI’s inaugural developer conference, DevDay, the GPT Store was announced but postponed until December, perhaps owing to a leadership change in November. (OpenAI’s board of directors forced CEO Sam Altman out before reinstating him with a new board after investors and staff panicked.)
Developers may create basic or complicated GPTs without coding. A GPT may be trained on a cookbook collection to answer recipe ingredient inquiries. Developers might use a GPT to evaluate their style or produce best-practice code from a company’s private codebases.
Developers may enter their GPT’s capabilities in plain language, and OpenAI’s GPT Builder will try to create an AI-powered chatbot. Since shortly after DevDay, developers can directly create and distribute GPTs on ChatGPT but not publicly list them.
Whether the GPT Store will begin a revenue-sharing system is unknown. Altman and CTO Mira Murati informed my colleague Devin Coldewey in November that GPT monetization was undetermined, and the GPT Store’s launch email doesn’t mention developer compensation.
A spokesman for OpenAI told Eltrys more will be published next week.
As I wrote for Eltrys’s semiregular AI newsletter, OpenAI’s transformation from AI model supplier to platform has been fascinating but not unexpected. The launch of ChatGPT plug-ins in March introduced third parties into OpenAI’s model ecosystem for the first time, revealing the startup’s aspirations.
GPTs democratize generative AI app production, at least for OpenAI models. GPTs might destroy consultancies that develop GPTs for consumers.