Peter Steinberger, creator of OpenClaw — the AI assistant powering the agent-only social network Moltbook — is advocating for specialized AI over the pursuit of artificial general intelligence (AGI). In a recent appearance on the Y Combinator podcast, Steinberger argued that the most effective AI systems are not broad, all-encompassing intelligences, but rather tools finely tuned for specific tasks. While Silicon Valley’s dominant narrative often centers on AGI — a hypothetical system capable of human-level reasoning across nearly any domain — and even superintelligence, which would surpass human cognition, Steinberger believes this vision is misguided. “What can one human being actually achieve?” he asked. “Do you think one human being could make an iPhone or go to space?” He pointed out that human progress stems from specialization: individuals and teams focus on narrow areas, and collectively, society achieves extraordinary results. The same principle, he argues, should apply to AI. Although current models are often labeled as “general,” in reality, they are already highly specialized — whether solving complex math problems, identifying genetic mutations, or analyzing medical images. The idea of a single, universal AI remains theoretical, while practical breakthroughs are coming from domain-specific systems. Startups and tech giants are increasingly embracing this approach. Axiom, founded by former Meta researcher Carina Hong and backed by $64 million in seed funding, is building AI focused on advanced mathematics. Google DeepMind has developed AlphaGenome, a system designed to predict how DNA mutations influence gene regulation. These efforts reflect a growing shift toward AI that excels in well-defined fields rather than attempting to master everything. Smaller, more efficient models are also gaining traction. Aidan Gomez, CEO of Cohere, told Business Insider in 2024 that the industry is under increasing pressure to prioritize model efficiency and intelligence through better data and algorithms, rather than simply scaling up model size. Steinberger isn’t alone in challenging the AGI narrative. Timnit Gebru, a prominent computer scientist and founder of the Distributed AI Research Institute, has called AGI a “fictional thing” in a November video published by Nature. She argues that real progress comes from well-scoped, testable systems and warns that the pursuit of an undefined, all-powerful “machine god” leads to exploitative labor practices and significant environmental costs. Steinberger, previously known for founding PSPDFKit, a PDF processing company, returned from retirement to work on AI. His journey reflects a broader evolution in the field — from tools that could simply automate typing to systems that now act as intelligent agents. His current mission: building AI that excels at specific tasks, not one that claims to do everything.