Raw LLM Responses

ChatGpt: Roman Yampolskiy (the University of Louisville computer scientist who writes a lot about AI safety) is deliberately provocative when he talks about timelines. His view sits on the very short end of the spectrum. A couple of things to keep in mind about why he says “just a few years” and ties that to massive unemployment: Acceleration argument ⚡ He points to how fast AI capabilities have jumped in just the past few years (from GPT-2 to GPT-4, from clunky image generators to photorealistic video). If progress continues exponentially, then a system smarter than humans in most domains could appear sooner rather than later. Task automation vs. job automation 👷 He often warns that once AI can perform all economically valuable tasks more cheaply and reliably than humans, employment as we know it collapses. That’s where his “99% unemployment” projection comes from — not because every job instantly vanishes, but because AI can outcompete humans at scale. Safety and control worries 🚨 Yampolskiy is one of the people who stress that superintelligence isn’t just about convenience or efficiency; it’s about whether we can keep control once we build something smarter than us. Of course, not all experts agree with him. Many argue that: Technology doesn’t diffuse instantly — adoption is gradual, and industries adapt. New jobs and roles emerge even as old ones disappear. Predicting “years away” has been a recurring theme in AI for decades, and the field is littered with over-optimistic forecasts. So Yampolskiy is basically ringing the alarm bell — “if this is coming quickly, society isn’t ready.” Whether you take his timeline literally or more as a call to prepare depends on how you read the speed of current breakthroughs.