In a bold vision for the future of artificial intelligence, Microsoft's AI CEO, Mustafa Suleyman, has outlined a crucial new benchmark. He calls it the test for Artificial Capable Intelligence (ACI), a significant stepping stone on the path to the holy grail of AI: Artificial General Intelligence (AGI).
The Modern Turing Test: From $100k to $1 Million
Suleyman's proposed test is deceptively simple yet profoundly challenging. He asks whether an AI agent can take a seed capital of $100,000 and legally grow it into $1 million. He described this as the "modern Turing Test," a measure that would demonstrate an AI's ability to plan, reason, make complex decisions, and operate effectively within real-world legal and financial frameworks.
"The next big milestone I'm watching for on our way to AGI: Artificial Capable Intelligence (ACI). Can an agent take $100k and legally turn it into $1M? To me that's the modern Turing Test," Suleyman wrote in a post on X, formerly known as Twitter.
The Industry's Big Bet on Autonomous AI Agents
Suleyman's comments arrive amidst a massive industry push towards developing sophisticated AI agents. Tech leaders like OpenAI CEO Sam Altman and Salesforce CEO Marc Benioff are heavily investing in this technology, seeing it as the next major revolution in enterprise software.
Companies including Microsoft, OpenAI, and Salesforce are actively promoting agents designed to execute tasks—from writing software code to managing intricate business workflows—with minimal human intervention. Last year, Marc Benioff hailed this shift as a "digital labor revolution," noting that Salesforce's AI technology had achieved about 93% accuracy and was handling up to 50% of the work within the company.
Adding to the optimistic forecasts, Sam Altman predicted in 2025 that AI could soon automate up to 40% of current human tasks. He has also suggested that AGI could become a reality before 2030.
Dissenting Voices and the "Slop" Critique
However, not all AI pioneers share this unbridled optimism. OpenAI co-founder and renowned AI researcher Andrej Karpathy has recently pushed back against the intense hype surrounding "agentic AI." He characterized the current generation of autonomous AI systems as "slop," suggesting the industry is overstating their capabilities.
"I feel like the industry is making too big of a jump and is trying to pretend like this is amazing, and it’s not. It’s slop," Karpathy stated on The Dwarkesh Podcast. He implied that some of the bullish timelines might be influenced by factors like fundraising, acknowledging that the technology is still at an intermediate stage of development.
The debate between Suleyman's ambitious milestone, the bullish predictions of Altman and Benioff, and Karpathy's cautious realism highlights the dynamic and uncertain frontier of AI development. The race is on to see which vision of the future—and which definition of capable intelligence—will ultimately prevail.
