Why AI Won't Take Over the World Anytime Soon
AI is smart but the world is slow—here's why your job is safe for now.
When AI like Claude Code started writing software, some experts feared a job apocalypse. But that hasn't happened. Why? Because the world has many steps, and making one step super fast doesn't speed up the whole process. Imagine a factory line: if one worker gets 100 times faster but others stay the same, the overall output barely improves. This is called Amdahl's law.
excels at tasks with clear rules, like math and coding. But real-world problems—like curing diseases or running a company—involve many bottlenecks: regulations, physical experiments, human coordination. Even if AI becomes infinitely smart at one step, other steps still hold things back. So we see constant improvements, not explosive leaps.
The author, once worried about rapid AI-driven job loss, now sees that predictions of economic shockwaves were wrong. Despite AI beating humans on benchmarks, millions still work in software. The lesson: intelligence isn't the only bottleneck. When AI becomes free and instant, other limits—like time, resources, and human factors—become the new bottlenecks.
This doesn't mean AI won't change things, but the change will be slower than some fear. The world is 'far too slow' for the kind of seismic transformation predicted by some AI enthusiasts. So for now, your job is probably safe—but it's worth watching how AI reshapes specific tasks.
- Amdahl's law: speeding up one part of a process doesn't speed up the whole if other parts are slow.
- AI is great at coding and math, but real-world progress is limited by many other factors.
- Predictions of mass unemployment from AI have not come true yet, and the economy changes slowly.
Why It Matters
AI won't instantly change everything—your job and daily life are likely safe for now, but stay informed.