OpenAI CEO hints at ‘some very good releases’ later this year
OpenAI CEO Sam Altman and several other business leaders 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 company’s products, such as ChatGPT, or general questions about artificial intelligence (AI) and artificial general intelligence (AGI). During the session, Altman said GPT-5 won’t be released this year, but the company plans to introduce “some very good releases” before the end of 2024.
OpenAI Staff hosts 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 Q&A post. The official X-handle (formerly known as Twitter) of OpenAI as well posted about the Reddit AMA.
Answering a question about the timeline for GPT-5 or its equivalent, Altman said: “We have some really good releases coming later this year! But nothing that we are going to call gpt-5.” This seems similar to 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 he finds it a faster and easier way to get to the information. He also emphasized that the web search functionality will be more useful for complex research. “I also look forward to a future where a search can dynamically display a customized web page,” he added.
Weil also answered several user questions. One asked about the delay in Sora, to which the OpenAI CPO said the delay was caused by the extra time needed to perfect the model, get security and imitation right, and the need to scale computing power. However, he did not announce a date for the launch.
Weil also emphasized that the ‘o’ series AI models, such as GPT-4o and o1-preview, will become a mainstay in the company’s lineup and will appear even after the release of GPT-5. Additionally, he also revealed that the ChatGPT Advanced Voice Mode could be customized to add a singing voice to the AI.
One user suggested: “Can we please get a hands-free way to end a voice call with ChatGPT? Use the shortcut for the action buttons on iPhone, but voice chat must be manually turned off each time. Weil seemed impressed by the idea, saying, “I love this idea. I’m now sharing it with the team!”
OpenAI SVP of Research Mike Chen also answered an important user question about AI hallucination. He explained why hallucinations from AI models haven’t completely disappeared, calling it a fundamentally difficult problem. This is because the AI models learn from human-written text, and humans can often make mistakes, which are then added to the core dataset of the large language models (LLMs).
“Our models are getting better at citing, which bases their answers on trusted sources, and we also believe that RL will also help with hallucinations – if we can programmatically check whether models are hallucinating, we can reward them if they don’t,” Chen added. .
