AGI (Artificial General Intelligence)
AGI is a hypothetical AI able to handle any intellectual task at human level, not only the ones it was trained for. No such system exists, and there is no agreed definition by which one would be recognised.
Why it is hard to discuss
The definitions differ fundamentally. For some, AGI is a system that passes any human test; for others, one that sets its own goals; for others still, an economic threshold such as replacing most remote work. These definitions give different answers to how close it is.
Current models are strong where data is plentiful and unexpectedly weak on simple tasks needing sustained step-by-step reasoning or physical intuition. Passing professional exams and working reliably on a real task are different capabilities.
Timeline forecasts differ by decades and come from people with a stake in the answer. That does not make them insincere, but it does mean specific dates deserve caution.
How to read claims about AGI
- Ask for the definition"AGI is close" without a stated criterion contains no checkable claim.
- A benchmark is not the jobA test score shows a test score. Test items frequently end up in training data.
- Watch the failures, not the winsWhere a system fails says more about how it works than a list of what it managed.
- The speaker has an interestTimeline estimates most often come from companies raising money against that goal.