Sam Altman, CEO of OpenAI, outlined a tantalizing glimpse of what people might expect from the long-awaited GPT-5 AI model. He tried to balance optimism and caution in his comments, but his view of the new model’s potential underscored his confidence that GPT-5 will represent a substantial improvement over its predecessor, GPT-4, and will not face any unsolvable problems.
“I hope it's a breakthrough,” Altman said. “A lot of the things that GPT-4 gets wrong, you know, they can't do much in the way of reasoning, sometimes they just go completely off track and make a silly mistake, which even a six-year-old wouldn't make.”
Altman compared the current state of AI technology to the early days of the iPhone, suggesting that while current models are useful, they are still in the early stages of realizing their potential. He noted that current AI models, including GPT-5, are relatively small compared to what future advancements could bring.
Interestingly, Altman's recent comments on model size indicate a slight shift from his previous stance. For those following Altman's comments closely, it's a sharp turnaround from when he suggested the era of giant models might be coming to an end last year. Instead, he now apparently believes models will likely continue to grow, driven by significant investments in computing power and energy.
Altman is confident that GPT-5 will address many of GPT-4's shortcomings, particularly in areas such as reasoning and error prevention. But Altman also stressed that while GPT-5's development is promising, there is still much work to be done. “We don't know yet. We're optimistic, but we still have a lot of work to do.”
See in the
The big picture for large language models
Altman addressed some of the biggest controversies around AI, particularly content licensing. He took the opportunity to boast about OpenAI’s approach, which involves deals with publishers to license news content to ChatGPT in exchange for training data for the models. He contrasted this approach with that of companies like Google, which claims that AI-driven traffic benefits publishers — a claim he and many others view with skepticism.
Altman also tempered expectations during the interview about what AI means for the Internet and the broader economy. He simultaneously suggested that there won't be a massive impact on Internet usage, while pushing for new approaches to commerce.
“I think AI won't change in a very significant way, but it will significantly change the way people use the internet,” Altman said. “And if it does, then some of the economic models of the past may have to evolve, and I think that's a bigger conversation than just training data.”
Altman suggested that GPT-5 is just the beginning of a series of advances aimed at building more sophisticated and capable AI systems. The next few months will be crucial in determining whether GPT-5 can deliver on its promise of a giant leap forward, addressing the limitations of its predecessors and paving the way for more advanced AI applications.