A Quiet Skeptic in the AI Gold Rush: Why One Engineer Says True Machine Intelligence Isn't Close

In 2026, the belief that artificial general intelligence (AGI) is nearly here has become an article of faith in tech and finance, fueling unprecedented investment. Yet, a dissenting voice from within the field is challenging this consensus with a detailed technical argument.

Software engineer and researcher Daniel Lants has published an essay titled “AGI Is Not Imminent.” His position isn't that today's AI lacks value, but that the core strategy of simply making models bigger is hitting a wall. According to Lants, scaling up current transformer models, while creating powerful tools, won't bridge the gap to the flexible, reasoning intelligence that defines AGI. This puts him at odds with prominent industry leaders who have publicly speculated about AGI arriving within a few years.

Lants draws a sharp line between what systems like large language models do—expert pattern-matching within their training data—and what true general intelligence requires. AGI would need to handle novel situations, conduct original research, and reason causally. Current models, he argues, cannot do that, no matter how much more computing power they get.

This skepticism extends to how progress is measured. Lants points out that high scores on professional exams or other benchmarks can be misleading. Models can learn to game these tests without possessing real understanding, often failing at simple tasks that require genuine logic.

The economic stakes are enormous. Billions are being spent on data centers and chips based on the expectation of relentless, scaling-driven progress toward AGI. Lants suggests the investment thesis may be flawed. He clarifies that useful, profitable AI tools are here now, but conflating them with imminent AGI is a dangerous mistake.

Lants doesn't claim AGI will never happen. His call is for intellectual honesty about the profound technical hurdles that remain, which likely demand entirely new approaches. As capital continues to flow, his analysis serves as a necessary counterweight to the prevailing optimism, reminding the industry that between today's impressive tools and tomorrow's promised general intelligence lies a chasm we don't yet know how to cross.