OpenAI CEO forecasts capable GPT-5, artificial GenAI

CALIFORNIA, UNITED STATES — Sam Altman, CEO of artificial intelligence (AI) research company OpenAI, reportedly offered an optimistic outlook on the future capabilities of large language models like GPT-5 and the potential for artificial general intelligence (AGI) during a talk with Y Combinator founders.
According to YC founder Richard He, as shared by AI/Data Executive Howie Xu on X (formerly Twitter), Altman encouraged them to build their companies with the mindset that GPT-5 and AGI will be achieved “relatively soon.”
Altman also stated that “most GPT-4 limitations will get fixed in GPT-5.”
At @ycombinator W24 kickoff today, @sama suggested ppl build w/ the mindset GPT-5 and AGI will be achieved "relatively soon"; most GPT-4 limitations will get fixed in GPT-5, per YC founder Richard He.
Expect @OpenAI #GPT5 in 2024 and #AGI in 2025?
What's your interpretation? 😂 pic.twitter.com/jehDupePLe— Howie Xu (@H0wie_Xu) January 12, 2024
Outsourcing advisor John Walter advised customer experience industries to closely monitor such major AI advances that can reduce common issues like hallucination.
“We saw a significant reduction in hallucination between GPT-3.5 and GPT-4. And we should expect that trend to continue with GPT-5,” said Walter in a LinkedIn post.
“It’s possible that GPT-5 will reduce hallucination enough to allow enterprise companies to confidently allow their customers to interact directly with generative AI chatbots and voice assistants.”
While details remain scarce, Altman’s remarks indicate active development of GPT-5, likely building on GPT-4’s foundation.
During a recent podcast with Bill Gates, Altman discussed the essential areas of development for GPT-5 including:
- Multi-modality – Integrating various forms of input and output, like speech and images to enhance AI’s interaction capabilities with humans.
- Reasoning and Reliability – To ensure that the AI consistently provides the best possible response out of thousands of possible answers.
- Customizability and Personalization – Catering to diverse user needs including adapting to individual preferences and styles.
- Personal Data Integration – Using personal data more effectively, such as integrating with email, calendars, and other external data sources.