OpenAI chief executive Sam Altman is scheduled to speak with CNBC before the company’s DevDay keynote, according to the broadcaster’s preview of its coverage. The programme also includes forthcoming announcements aimed at developers.

CNBC’s preview establishes the planned interview and event coverage, but provides no specific product details, release terms or remarks from Altman. It does not identify which developer capabilities will be announced or when any resulting products might become available. Those details remain unconfirmed in the supplied reporting.

DevDay is OpenAI’s developer event. Its audience includes people building software around artificial intelligence, a different use case from interacting directly with a consumer chatbot. A developer might incorporate an AI model into a customer support tool, a document processing system or an application that helps users write and revise material.

The connection between a software application and a model is often an application programming interface, or API. The application sends a request to a service and receives a response it can display or process. Developers write the surrounding software that decides what information to send, how to handle the answer and what happens if a request fails.

That distinction helps explain the technical focus of developer announcements. A model’s abilities are one component of an application; access methods, usage limits, documentation and operating terms are separate components. The preview does not establish whether OpenAI plans to change any of them.

How developer releases reach users

A presentation, a demonstration and a usable release describe different stages of a product’s introduction. A demonstration shows a capability under selected conditions. A preview can give a limited group access for testing. General availability typically means a product is offered more broadly under published conditions, though eligibility and restrictions can still apply.

For software builders, documentation supplies the instructions needed to turn an announcement into an implementation. It describes how requests must be structured, what responses contain and which limitations apply. Sample code can illustrate an intended workflow, while release notes explain changes that may affect existing integrations.

Pricing is another distinct part of the mechanism. AI services can charge according to usage, including the volume of information processed or generated. Text models commonly measure that volume in tokens, units that can represent words, parts of words or punctuation. A quoted token price therefore cannot be read directly as a fixed price per question: requests and answers vary in length.

Testing also extends beyond whether a system can produce an impressive answer. Developers evaluate response time, consistency, error handling and performance on tasks representative of their own users. Applications handling personal or business information also require an understanding of the service’s data handling terms. These are general considerations for AI software development; CNBC’s preview provides no new OpenAI terms on these subjects.

What to watch

The next reporting checkpoints are Altman’s interview, the keynote and the developer announcements. Confirmed documentation, availability dates and access conditions would establish what builders can actually adopt. Until those details emerge, the supplied reporting supports the event’s planned coverage, without establishing the substance of its releases.