AI has made extraordinary progress over the last three years, but those at the center of these developments now seem to believe that this progress is set to accelerate even further.
OpenAI CEO Sam Altman is rarely short of bold predictions, but his latest remarks carry a weight that goes beyond typical Silicon Valley optimism. In a striking admission, Altman has revealed that even he — one of the most prominent figures in the AI industry — was caught off guard by how quickly the technology is advancing.
Altman was asked about the resignation of an Anthropic employee who said that they were apprehensive about the direction AI was taking and wanted to simply go off to England to write poetry. “The inside view at the companies looking at what’s going to happen — the world is not prepared,” Altman replied. “We’re going to have extremely capable models soon. It’s going to be a faster takeoff than I originally thought. And that is stressful and anxiety inducing.”
The candor is notable. Altman has long been an evangelist for AI’s transformative potential, but framing the pace of progress as something that causes him personal stress and anxiety signals a shift in tone — one that the broader business world would do well to pay attention to.
The phrase “faster takeoff” is significant in the context of AI development. In the field, “takeoff” refers to the speed at which AI systems transition from their current capabilities to something far more powerful. A fast takeoff implies that the window for adaptation — for businesses, governments, regulators, and society at large — could be much narrower than previously assumed. If Altman, who has access to the most advanced models in development at OpenAI, is recalibrating his own expectations upward, that is a meaningful data point.
There are indications that this is happening. ARC-AGI, a benchmark which was meant to test AI models on whether they could have human-like intelligence, is already saturated. The researchers came up with an ARC-AGI 2 benchmark, but just yesterday, Google managed to double its performance on it with a 77% score. The pace of development across other benchmarks across coding, long task performance and research also seems to be accelerating.
For the business community, the implications are hard to overstate. Companies that have been taking a measured, wait-and-see approach to AI integration may find that timeline is a luxury they can no longer afford. Industries from finance and healthcare to logistics and legal services are already contending with AI tools that are reshaping workflows and competitive dynamics. A faster-than-expected acceleration would compress the time available to adapt strategies, retrain workforces, and rethink entire business models.
Altman’s remarks also raise pointed questions for policymakers. Regulatory frameworks for AI are still in their infancy in most parts of the world, with the European Union’s AI Act representing one of the most advanced attempts to govern the technology. If even the CEO of the world’s most prominent AI company believes the world is unprepared for what is coming, the urgency for thoughtful, informed regulation has never been clearer.
There is, of course, a complexity to Altman’s position. As the leader of the organization arguably most responsible for driving this acceleration, his warnings exist in tension with his role. OpenAI continues to push the frontier forward at pace, even as its CEO acknowledges the anxiety that progress produces. Critics have long pointed to this tension as a fundamental contradiction at the heart of the company’s stated mission of ensuring AI benefits humanity.
But whatever one makes of that tension, the underlying message is one the world should take seriously. The people building these systems — with the clearest view of what is coming next — are telling us to expect more, sooner. The question now is whether businesses, governments, and institutions can move fast enough to meet that reality.