I never got this. Why not just run the browser in a VM, capture the screen and do a mouse click. I mean, why bother with all the headless nonsense which is an

I never got this. Why not just run the browser in a VM, capture the screen and do a mouse click. I mean, why bother with all the headless nonsense which is an arms race. How could google possible ever defeat something that never ever even goes near the operating environment and just appears as a simple human mouse click? I suppose you’d have to simulate the human movement of the mouse over to make it look like a human actually did it, but how hard can that be? Just train it with a few 100 Turks moving mouse pointers to click on links. Though an interesting counter measure might be to inject cpu spikes and measure the impact in the mouse movements and the robot controlling it. Must be a fun job, both sides.