I was playing a racing game on my phone the other day. Every time I pick up a new game, I am amazed at how things look and how incredibly well they play on a device like the iPhone. I am blown away at how far technology has come. It’s likely I’m more easily impressed than the average person because my last real console was the Sega Genesis back in high school and I briefly flirted with the Playstation Portable a while back. (We broke up because, well, startups.)

When I play this racing game, the only way I can win against the stronger drivers is to go into the turns as aggressively and as fast as I can and then to force myself into the inside corner. I pull this move on every single corner in the game. I do it even if it means going over the rumble strip and side-swiping other cars and losing my sideview mirror and risking a spin into the wall. After all, I can play dangerously because both “my health” in the game and the car’s damage can be easily fixed at the end of the race.

It got me thinking: were we ever to build a “perfect” driving simulator, right down to the physics and the weather and the response of the tires, would a professional race car driver be able to drive faster records in the simulator than in a real-life car? It’s not so much that he is afraid to bust up the car – at the high end of racing, I imagine that doesn’t matter so much because there’s enough money to fix everything or even get a new car. It’s not so much that he will be more tired during a real race than a virtual race (he would be equally strained in the perfect virtual clone). But the real reason he might do better is that there’s no fear in a simulator.

A couple of days later, I came across this report of an driverless Audi race car that hit 190 mph when racing a human driver – and it beat the human driver by five seconds! There’s faster decision-making inside the computer around the physics of racing and the track, of course – and there’s the element of tirelessness: less fatigue in the driverless car than in a real driver. But again, I wondered if its fearlessness that really matters here: the machine knows no such thing as fear.

Then, earlier this week, another intelligent read on machines taking a human’s place: how a boy with autism found friendship and someone to talk to in Siri. Machine-as-sidekick is powerful because of its patience in being able to hold a conversation and a presence wherever required far longer than a human. “Getting results requires a lot of repetition. Humans are not patient. Machines are very, very patient.”

So perhaps, when it comes to boundaries, it’s not so much what the machines know that we don’t – but rather what they don’t know that really enables them to surpass us.

If you don’t know where the boundaries are, you push yourself just a little bit more than an observer expects. If you don’t know where the boundaries are, there are no boundaries.