Moravec Revisits His 1988 Forecast As AI Deadline Nears

Hans Moravec made his call in 1988: forty years to human-level machine intelligence. That puts the deadline inside the current decade, and New York magazine's Intelligencer has gone back to the Carnegie Mellon roboticist to ask what he thinks now, as machines climb what the magazine describes as the path he mapped.

The framing in the piece is that Moravec was right about AI, and that the harder open question is whether he is also right about what comes next for humans. His long-standing answer, dating to his 1988 book Mind Children, is that intelligent machines are best understood as our evolutionary heirs rather than our rivals. That is a philosophical bet, not a technical one, and it is a much longer walk from a compute-and-capability forecast to a claim about the fate of a species.

Why this matters if you are not a longtime Moravec reader: a lot of the current AI discourse treats near-term capability timelines as if they were invented last year. They were not. Forecasts made when the field was small and unfashionable are now being scored against real systems, and the ones that leaned on raw compute growth have aged unusually well. That is a data point about which kinds of predictions to take seriously in the next round, and which to keep at arm's length.

The honest caveat is that 'he was right' flattens a lot. Human-level at what tasks, measured how, on which hardware, is precisely the argument the field is still having, and the outlet's own promotional framing does not settle it. What the reporting also does not appear to give you, at least from what is public around the piece, is a fresh technical call from Moravec on the language-model-plus-robotics blend that actually arrived, versus the more embodied path he originally emphasized.

The forward-looking part is worth sitting with. If you take the timeline call seriously, the interesting move is not to argue about the deadline but to pressure-test the successor thesis before it hardens into common sense. That is the conversation a piece like this can start.