Artificial General Intelligence (AGI) may already exist, according to Eliza Labs founder Shaw Walters. Speaking at ETHDenver, Walters asserted that current leading AI models fulfill his definition of AGI.
"I think that we're at the inflection point where we have AGI," Walters stated. "I completely believe that this is general intelligence. It's nothing like us. It learns in a completely different way, but it is intelligent nonetheless, and it is very general."
Eliza Labs, originally launched as ai16z, developed the open-source ElizaOS, a framework for creating autonomous AI agents for blockchains. AGI, a term popularized in the late 1990s, refers to AI capable of matching or exceeding human cognitive abilities across a broad range of tasks.
While many AI leaders anticipate AGI's arrival within the next decade, Walters rejects the notion of a single, dominant system, stating, "I just do not see it as the AI God. There's never going to be one, because life loves variants."
Walters detailed his work on AI agents, noting significant progress with GPT-4's release in 2023, which enabled more reliable structured responses and action calling. This marked a shift from experimental chatbots to persistent systems embedded across various platforms.
However, the increasing autonomy and wallet control of AI agents have introduced significant security risks. Walters cautioned that as AI advances toward AGI, it becomes less predictable, akin to a fallible human rather than a calculator, making foolproof safeguards challenging to engineer. "At the end of the day, you're dealing with something that's more like a human and less like a calculator," he said. "It's gonna do stupid things sometimes, and there’s just no way to build a super secure system that's going to keep them from doing something dumb."