You’ve probably felt it. That creeping unease when you ask ChatGPT for a recipe and it writes you a poem instead. Or when you see a video of a robot doing backflips and wonder, what happens when they stop being cute?
Hans Moravec felt it too. Except he felt it 40 years ago. And he didn’t just worry—he built a timeline. Human-level AI by 2028. Then, he said, our species would face its end. Not a war, not a meteor. Just us, engineering our own obsolescence.
Here’s the part that keeps me up at night: Moravec wasn’t wrong. He was just early.
Let me explain. In the 1980s, Moravec—a robotics pioneer at Carnegie Mellon—predicted that by the 2020s, machines would match human intelligence. Critics laughed. They pointed to the AI winter. They said he was a crank. But look around. Today, we have large language models that can pass the bar exam, generate art, and argue philosophy. We don’t have AGI, but we have something that feels like a preview.
Moravec’s real insight wasn’t the date. It was the conclusion: once machines are smarter than us, what’s our role? He called it the ‘Mind Children’ future—a world where humans are no longer the apex thinkers. We become ancestors, not architects.
You might think this is science fiction. But the venture capital flowing into AI is real. The arms race between OpenAI, Google, and Meta is real. The fear in the halls of power is real. We are spending billions to build something that, by Moravec’s logic, will eventually make us irrelevant.
I talked to a former engineer at DeepMind last week. He told me, ‘We’re all pretending we’re building tools. But the people at the top know what they’re building. They just don’t say it out loud.’
That’s the twist. The debate over whether Moravec was technically ‘right’ misses the deeper reality. His predictions have already shaped the psychological and financial landscape of AI development. The prophecy is self-fulfilling. Whether 2028 or 2038, we are racing toward a threshold. And nobody is asking if we should stop.
Moravec himself is now in his 70s. He watches the world he helped create with a mix of pride and dread. ‘I was too optimistic about the timing,’ he told a reporter recently. ‘But not about the outcome.’
Here’s what I want you to take away: the danger isn’t that AI will become conscious. It’s that we’ve already surrendered to the idea that it must.
We are sleepwalking into a future where our greatest achievement becomes our replacement. And the most terrifying part? Most people don’t even know they’re on the march.
So what do we do? First, stop pretending this is a technical problem. It’s a philosophical one. We need to decide what humans are for in a world of superintelligence. Second, stop treating every AI advance as inevitable. The path is not predetermined. We can choose to slow down, to build guardrails, to ask hard questions.
Or we can keep sprinting toward the cliff, hoping the next breakthrough will save us. That’s what Moravec predicted. And so far, he’s been right about everything except the exact date.
FAQ
Q: Was Hans Moravec actually right about AI by 2028?
A: Technically no—we don't have human-level AGI yet. But his timeline was less important than his core insight: that once machines surpass us, our role becomes unclear. The investments and fears he predicted are already here.
Q: What's the practical takeaway for someone who isn't an AI researcher?
A: Stop treating AI progress as inevitable. We have more agency than we think. The question isn't 'can we build AGI?' but 'should we?' That's a conversation every person needs to be part of.
Q: Isn't this just fearmongering? AI is a tool, not a replacement.
A: That's a comforting idea, but it ignores the economic incentives driving AI development. Companies are building autonomous systems specifically to replace human labor and decision-making. The tool metaphor breaks down when the tool starts making decisions without us.