Is Artificial General Intelligence Already Here? One AI Founder Thinks So
Shaw Walters, founder of Eliza Labs, claims that artificial general intelligence (AGI) has arrived, asserting current AI models already display general intelligence, though different from human cognition. Walters emphasizes these AIs are intelligent and broadly capable, but learn differently from humans. He rejects the idea that AGI will emerge as a single, “god-like” system, arguing that diversity in intelligent systems is inevitable. Walters’ work began during the GPT-3 era, focusing on improving agent reliability, with a significant leap after GPT-4’s release, which enabled more structured and actionable outputs. Progress in AI agents has led to their widespread integration in crypto and consumer platforms, but Walters notes this evolution has heightened security concerns, especially as agents gain extensive access and autonomy. He warns that as AI approaches AGI, it becomes unpredictable and more akin to fallible humans rather than calculators, making perfect safeguards unattainable. Walters concludes that AI agents will inevitably make mistakes, and absolute security is unachievable as these systems continue to advance.
