The economy doesn't need true AGI, says Replit CEO

""We can get to like functional AGI," Masad said. He defines functional AGI as AI that doesn't need human-like consciousness or reasoning, just systems capable of learning from real-world data and completing verifiable tasks on their own. "We'll target every sector economy and you can automate a big part of labour that way," he said. "We're on that track for sure.""

"Forget building a god-tier superintelligence. Masad said he's not convinced that we'll ever reach true AGI, the kind of artificial intelligence that can learn and adapt across knowledge fields like a human mind. While true AGI could "propel us to the next level of human civilization," Masad said he's "bearish on true AGI breakthrough because what we built is so useful and economically valuable.""

"While Silicon Valley obsesses over true AGI, the practical version is already within reach, and it's good enough to automate massive chunks of the economy. Masad also said the industry might be stuck in a "local maximum trap," which means AI companies are optimizing what already works instead of reinventing the field. "Maybe the general problem is actually not within our lifetimes," Masad said, referring to solving the problem of general intelligence itself. "Who knows?""

Functional AGI is defined as AI that does not require human-like consciousness or reasoning and instead learns from real-world data to complete verifiable tasks autonomously. Such systems can be applied across every economic sector and have the potential to automate large portions of labor. True AGI that can learn and adapt across knowledge fields like a human mind remains uncertain and may not be achieved soon. The industry risks a local-maximum trap by optimizing existing large-model approaches instead of pursuing radical new paradigms. The current practical value of AI reduces incentives for seeking a true AGI breakthrough, making its arrival uncertain.

Read at Business Insider