Marc Benioff, CEO of Salesforce, believes artificial intelligence (AI) is nearing the “upper limits” of what large language models (LLMs) like those powering ChatGPT can achieve, Business Insider reports.
Speaking on The Wall Street Journal’s “Future of Everything” podcast, Benioff emphasized that the next wave of AI innovation lies in autonomous agents — systems capable of independently performing tasks.
“We’ve all been caught up in the excitement around ChatGPT,” Benioff said.
He suggested that LLMs have led to an overestimation of AI’s current capabilities. While LLMs have been revolutionary in natural language processing and conversational AI, Benioff argued they are not the ultimate solution for advancing AI. Instead, autonomous agents — AI-driven systems that can execute tasks like sales communications or managing marketing campaigns — hold greater promise for transforming businesses.
Salesforce has been at the forefront of this shift, offering AI agents that automate customer service and other enterprise tasks. Other tech leaders, such as OpenAI, are also developing agents that perform specific functions, including coding and travel booking.
Benioff cautioned against unrealistic portrayals of AI capabilities, often fueled by popular culture. He referenced films like Terminator (1984), Minority Report (2002), and WarGames (1983) as examples of how public imagination has outpaced current technology.
“We are not at the point where AI can predict crimes or launch nuclear weapons like in these movies… Yes, AI can improve productivity, customer relationships, and business margins. But the idea that it can solve existential problems like curing cancer or climate change is misleading,” Benioff said.
Benioff’s perspective aligns with comments from Nvidia CEO Jensen Huang, who recently predicted a future where employees work alongside AI agents to enhance productivity and efficiency. Benioff underscored this, highlighting the role of AI in improving business processes without succumbing to the overblown narratives propagated by some AI evangelists.
“We are in an era where AI can augment human effort and deliver tangible benefits, but we are far from the science fiction scenarios depicted in the movies,” Benioff said.
As companies like Salesforce and OpenAI invest in autonomous agents, the broader tech industry is gearing up for a shift from LLM-driven AI toward systems designed to function independently. While acknowledging the transformative potential of these technologies, Benioff remains grounded about their current limitations and realistic applications.
