OpenAI Will Not Release GPT-5 This Year But ‘Some Very Good Releases’ Are Coming, Says CEO Sam Altman


OpenAI CEO Sam Altman and several other company executives hosted an ask-me-anything (AMA) session on Thursday. The session was hosted on the social networking platform Reddit and users were told to ask questions about the AI firm’s products such as ChatGPT or general queries about artificial intelligence (AI) and artificial general intelligence (AGI). During the session, Altman said that GPT-5 will not be released this year, however, the company plans to introduce “some very good releases” before the end of 2024.

OpenAI Staff Host AMA on Reddit

The AMA session was hosted on the ChatGPT subreddit. Calling it “our Reddit launch”, several executives including Altman, OpenAI CPO Kevin Weil, SVP of Research Mark Chen, VP Engineering ​​Srinivas Narayanan, and Chief Scientist Jakub Pachocki participated in the question-and-answer post. The official X (formerly known as Twitter) handle of OpenAI also posted about the Reddit AMA.

Answering a question about the timeline for GPT-5 or its equivalent’s release, Altman said, “We have some very good releases coming later this year! nothing that we are going to call gpt-5, though.” This seems on par with what multiple reports have confirmed with most expecting OpenAI to release the next flagship model sometime in 2025.

Another user asked about the value that SearchGPT or the ChatGPT Search feature brings, Altman said that he finds it to be a faster and easier way to get to the information. He also highlighted that the web search functionality will be more useful for complex research. “I also look forward to a future where a search query can dynamically render a custom web page in response,” he added.

Weil also fielded several questions from users. One asked about the delay in Sora, to which the OpenAI CPO said that the delay was caused due to additional time taken to perfect the model, getting safety and impersonation right, and the need to scale compute. However, he did not reveal a date for its launch.

Weil also highlighted that the ‘o’ series AI models, such as GPT-4o and o1-preview, will become a mainstay in the company’s lineup and will make an appearance even after the release of GPT-5. Additionally, he also revealed that the ChatGPT Advanced Voice Mode could be tweaked to add a singing voice to the AI.

One user suggested, “Can we please get a hands-free way to end a voice conversation with ChatGPT? Using the action button shortcut on iPhone, but the voice chat needs manually dismissed every time.” Weil seemed to be impressed with the idea and said, “I love this idea. Sharing with the team now!”

OpenAI SVP or Research Mike Chen also answered an important user question about AI hallucination. Explaining why hallucinations from AI models are not completely gone, he called it a fundamentally hard problem. This is because the AI models learn from human-written text, and humans can often make errors, which then are added to the core dataset of the large language models (LLMs).

“Our models are improving at citing, which grounds their answers in trusted sources, and we also believe that RL will help with hallucinations as well – when we can programmatically check whether models hallucinate, we can reward it for not doing so,” Chen added.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *