OpenAI Sam Altman finally knows when ChatGPT will get the GPT-5 upgrade, according to MIT Technology Review. He said “yes” when asked about it during an interview in Cambridge. But, of course, he didn’t reveal the actual release date of the highly-anticipated GPT-5 upgrade.
As a ChatGPT user who has been following GenAI developments closely, I know the GPT-5 is coming soon. Rumors say it’ll happen this year. But even without leaks, I’d still expect OpenAI to want to upgrade ChatGPT just so it can keep up with and pass Google. I say that as the latter’s Gemini 1.5 Pro update is expected to be much better than GPT-4.
But I’m not even that interested in ChatGPT getting better at reasoning or faster at spewing out answers to my prompts. What I really want from the whole AI revolution is a new computing experience where personal AI is readily available to me. And Altman just teased exactly that in the same chat with reporters.
“What you really want,” Altman told MIT Technology Review, “is just this thing that is off helping you.” Why yes, that’s exactly what I want from ChatGPT and its rivals.
Altman described the killer app for AI as a “super-competent colleague that knows absolutely everything about my whole life, every email, every conversation I’ve ever had, but doesn’t feel like an extension.” That gets another “yes” from this AI fan.
I’ve been saying for months that the next step in genAI is personal AI. Companies like Google and Apple are likelier to get us there before OpenAI and Anthropic. The latter two are the respective makers of ChatGPT and Claude, AI platforms that lack hardware.
I’ve also said that personal AI has to be secure, private, and accurate. That is an assistant I’d be ready to trust with personal data, knowing it’ll never hit a company’s servers or be used to train the models. I want an AI assistant that’s also mostly correct with its responses.
Altman explained that this “super-competent colleague” would be able to tackle some tasks instantly. It might have to ask you questions for the more complex ones if it fails to achieve the goal after a first attempt.
This sounds like Altman describing potential functionality coming to GPT-5 or future versions of ChatGPT. It might be just wishful thinking from yours truly. But I’ll remind you that OpenAI just gave ChatGPT the ability to save memories about the chat and the user. This key upgrade paves the way for ChatGPT to evolve into a personal AI assistant.
Altman also referred to GPT-4 as “incredibly dumb” in the interview, which is something he said before. But GPT-4 is incredibly dumb compared to what’s coming next to ChatGPT.
Anyone in the room with Altman making these claims would ask the obvious question: Will we need ChatGPT hardware for that sort of AI experience? After all, Altman is rumored to work with Jony Ive on the iPhone of artificial intelligence.
The CEO answered that question, saying that there’s a chance we won’t need a device at all. That such an advanced personal AI app could live in the cloud. But even if you won’t need a new device, “you’ll be happy to have” one, the CEO quickly added. At the same time, Altman says that new hardware is “far from my expertise,” as if he were already running ahead of potential rumors.
That’s fine, however. I’d expect Jony Ive to bring on the hardware-making expertise. Altman’s abilities would be needed elsewhere.
Getting back to my idea of personal AI, I’d love it if it ran on-device, on my current or future iPhone. While I wait, I’d definitely want to see OpenAI’s version of this “super-competent colleague that knows absolutely everything about my whole life, every email, every conversation I’ve ever had, but doesn’t feel like an extension” that Altman is teasing.