OpenAI chief executive Sam Altman said Thursday that the company is limited by computing capacity, which affects its ability to release products as frequently as desired.
When asked about the delay in publishing GPT-5, Altman answered that difficult decisions on compute allocation result from the complexity of the models and limited resources.
Reportedly needing help getting enough computational capability, the Microsoft (MSFT, Financials)-backed business has been developing custom processors with Broadcom (AVGO, Financials) and Taiwan Semiconductor Manufacturing Company (TSM, Financials). Though other releases are likely before the end of the year, Altman also pointed out that OpenAI will wait to release its next significant model this year.
Before moving on with a major update, Chief Product Officer Kevin Weil said the business is concentrated on perfecting the video model Sora, handling problems including safety, impersonation, and computation scaling. Altman also mentioned that although OpenAI's image model, DALL-E, has no planned release, the upcoming model will try to satisfy high expectations.
OpenAI lately refuted claims from The Verge on a scheduled Orion model release this year. Made during a "Ask Me Anything" session on Reddit following OpenAI's introduction of new ChatGPT search tools, which intensified Alphabet's (GOOGL, Financials) Google's competition, the remarks reflect.
OpenAI also covered recent high-profile executive departures, with VP of Engineering Srinivas Narayanan stressing the strength of the present team and recent new additions while acknowledging the absence of former colleagues.
