The AI Scaling Wall May Not Exist
Inside AI labs, researchers believe the most explosive phase of progress may still be ahead. For the past year, a common narrative has taken hold across the AI industry: Scaling is slowing down. Larger models are producing smaller gains. Benchmarks are improving more gradually. Some researchers have begun to argue that the explosive phase of large language models may already be behind us. But inside the labs building these systems, the story looks very different. At a recent Morgan Stanley Technology, Media & Telecom Conference, Anthropic CEO Dario Amodei dismissed the idea outright: “We do not see hitting the wall.” If anything, he suggested the opposite. The most dramatic phase of AI progress may still lie ahead — and it could arrive sooner than most people expect. If almost anyone else made that claim, it might sound like hype. But Anthropic sits at the center of the current race to build more capable AI systems. Its Claude models power a growing number of enterprise applications, and the company is widely considered one of the three or four organizations operating at the frontier of AI development. ...