Alexander Embiricos, head of product development for OpenAI’s Codex coding agent, identified human typing speed as a key bottleneck to achieving artificial general intelligence, or AGI. He made the comments on “Lenny’s Podcast” on Sunday.
Embiricos described human typing speed, or multi-tasking speed when writing prompts, as the “current underappreciated limiting factor” to AGI. AGI refers to AI that can reason as well as or better than humans, a goal pursued by major AI companies.
“You can have an agent watch all the work you’re doing, but if you don’t have the agent also validating its work, then you’re still bottlenecked on, like, can you go review all that code?” Embiricos said.
To overcome this, Embiricos called for redesigning systems to relieve humans of writing prompts and validating AI output. He argued that humans lack the speed for rapid progress in these tasks.
“If we can rebuild systems to let the agent be default useful, we’ll start unlocking hockey sticks,” he said. Hockey stick growth describes a pattern where progress remains flat before suddenly spiking upward.
Embiricos noted that no single path exists to fully automated workflows. Each use case will demand a tailored approach.
He predicted that starting next year, early adopters will experience sharp productivity increases, described as hockey stick gains. Over the following years, larger companies will achieve similar results.
AGI will emerge during the interval between early adopters’ initial productivity surges and the point when tech giants fully automate processes using AI agents, Embiricos stated.
“That hockey-sticking will be flowing back into the AI labs, and that’s when we’ll basically be at the AGI,” he said. This feedback loop of heightened productivity into AI research will drive the realization of AGI.
