What do I mean by “Artificial General Intelligence”?

According to LessWrong, researcher Steven Byrnes is drawing a distinction between the AI systems we have today and the artificial general intelligence that researchers worry about emerging in the years ahead. Today's AI requires massive, bespoke R&D for each new task—build a system that drives a car, and you'll need years and dozens of experts. Humans, by contrast, are born with a brain design barely changed in 100,000 years, yet can learn to drive, code, or make coffee with just a few hours of practice. That's the kind of flexible, general intelligence Byrnes is describing when he discusses AGI: a system that, like a human, could autonomously figure things out, make plans, pivot when things fail, and eventually tackle the kinds of scientific and technological breakthroughs that historically took entire societies centuries to achieve. Byrnes notes that we know AGI is possible because humans already do it—and since it doesn't rely on magic, just physics and math, future engineering should eventually replicate it. His real concern, though, isn't whether AGI will exist, but alignment: making sure that when it does, it wants to help humanity rather than replace us.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton