OpenAI CEO Sam Altman has addressed rumors concerning GPT-4 — the corporate’s as but unreleased language mannequin and newest within the GPT-series that varieties the inspiration of AI chatbot ChatGPT — saying that “people are begging to be disappointed and they will be.”
Throughout an interview with StrictlyVC, Altman was requested if GPT-4 will come out within the first quarter or half of the yr, as many anticipate. He responded by providing no sure timeframe. “It’ll come out at some point, when we are confident we can do it safely and responsibly,” he stated.
GPT-3 got here out in 2020, and an improved model, GPT 3.5, was used to create ChatGPT. The launch of GPT-4 is far anticipated, with extra excitable members of the AI neighborhood and Silicon Valley world already declaring it to be an enormous leap ahead. Making wild predictions in regards to the capabilities of GPT-4 has develop into one thing of a meme in these circles, significantly when it comes to guessing the mannequin’s variety of parameters (a metric that corresponds to an AI system’s complexity and, roughly, its functionality — however not in a linear vogue).
When requested about one viral (and factually incorrect) chart that purportedly compares the variety of parameters in GPT-3 (175 billion) to GPT-4 (100 trillion), Altman referred to as it “complete bullshit.”
“The GPT-4 rumor mill is a ridiculous thing. I don’t know where it all comes from,” stated the OpenAI CEO. “People are begging to be disappointed and they will be. The hype is just like… We don’t have an actual AGI and that’s sort of what’s expected of us.”
(AGI right here refers to “artificial general intelligence” — shorthand for an AI system with a minimum of human-equivalent capabilities throughout many domains.)
Video-generating AI fashions are coming, says Altman
Within the interview, Altman addressed numerous subjects, together with when OpenAI will construct an AI mannequin able to producing video. (Meta and Google have already demoed analysis on this space.) “It will come. I wouldn’t want to make a confident prediction about when,” stated Altman on generative video AI. “We’ll try to do it, other people will try to do it … It’s a legitimate research project. It could be pretty soon; it could take a while.”
The complete interview can be watched in two elements, right here and right here (with the second half focusing extra on OpenAI the corporate and AI extra typically), however we’ve picked out a few of Altman’s most notable statements under:
(This level is notable given ongoing conversations about AI and bias. Methods like ChatGPT have a tendency to regurgitate many social biases, like sexism and racism, which they internalize based mostly on their coaching knowledge. Corporations like OpenAI attempt to mitigate the biases by stopping the methods from repeating these concepts. Nonetheless, some conservative writers have accused ChatGPT of being “woke” due to its solutions to sure political and cultural questions.)
“Generated text is something we all need to adapt to, and that’s fine.”