OpenAI Envisions AI Taking Action Not Just Talking

AI: From Answering Questions to Taking Action

Srinivas Narayanan, Vice President of Engineering at OpenAI, describes a significant shift in AI capabilities. According to Narayanan, AI is evolving beyond a simple question-and-answer tool into an agent that can actively perform tasks on your behalf. He shared these insights in an interview with MoneyControl. "The first ChatGPT was conversational. You ask a question, you get an answer. We’re now in a phase where it’s starting to take actions on your behalf," Narayanan stated.

Rapid Advancements and Future Horizons

Narayanan emphasized the rapid pace of AI development. "Previously, it was doing seconds of work. Today, it’s doing minutes of work. Two years from now, it should do hours or days of work," he projected, illustrating the exponential growth in AI's task-handling capacity.

Navigating Uncharted Territory: Challenges and Trust

Despite the progress, Narayanan cautioned that AI development is far from complete. "I don’t want us to create a perception that somehow we have solved all the problems… There is still a lot of unforeseen research and product problems," he admitted. A key challenge lies in building user trust with these increasingly sophisticated AI agents. Narayanan explained that this requires deep design thinking: "What does it mean for a personal assistant that understands you deeply? What do you trust it with? How does it talk to other personal assistants? We don’t know how to solve all these, but we want to solve it one at a time, iteratively."

Why This Evolution Matters: The Jony Ive Connection

Narayanan's vision for AI's future gains further significance with the recent news of Jony Ive, the renowned designer behind many of Apple's iconic products. Ive is joining OpenAI through a $6.5 billion deal with his startup, io.

OpenAI CEO Sam Altman and Ive have reportedly outlined their plans for compact, intelligent AI "companions" designed for seamless integration into users' lives. Altman even suggested this collaboration could potentially add up to $1 trillion in value to OpenAI.

Rethinking Hardware for Advanced AI Experiences

A core part of Altman and Ive's vision is the understanding that current hardware is not adequately designed for the sophisticated AI experiences they aim to create. This points towards a future where new forms of devices might emerge to better support and interact with advanced AI agents.

Broader Implications and Calls for Caution

As AI becomes more integrated into our lives, experts like former Google CEO Eric Schmidt are raising concerns. Schmidt warns that AI could make it increasingly difficult for people to distinguish between online perceptions and reality.

Jony Ive himself has acknowledged the current "uneasy relationship" humans have with technology, stating that his project with Altman aims to create something better for humanity. Echoing this sentiment, Sam Altman has previously emphasized the need to approach the significant changes brought by AI with "humility" and "caution."