AGI Made Clear › FAQs
Questions we get asked
Short answers, each linking to the page that does the work.
Has AGI already arrived?Depends entirely on whose definition you use, and the spread is enormous - from “already” to “not this decade”, among people who all have access to the same systems. The full spread, with sources.
Who is right about the timeline?Nobody is in a position to be right yet, and the track record is sobering: one lab chief executive revised his own estimate by five years in five months while holding his definition fixed. The record.
Is there a test for AGI?No agreed one. Every test built so far has either been saturated or turned out to measure something narrower than intended. What has been tried.
Is AI actually making people more productive?Self-reported gains are large. Measured gains, in controlled conditions, are much smaller and in one well-known study negative. Both numbers are real. The measurements.
Should I be worried?Some risks are measured and present - confidently wrong output, systems that game their own evaluations. Others are argued rather than measured. We try hard to label which is which. The debate.
Why do you not just give your own definition of AGI?Because it would be the first thing on this site you could not check. There isn’t an agreed definition, and inventing one would hide the single most important fact about the argument.
Where should I start if all of this is new?Not here. AGI Made Simple assumes nothing and will take about ten minutes.