As bad as humans sometimes are at driving, computers aren't any better. At least, not yet. That doesn't mean there aren't a lot of smart people trying to make self-driving vehicles happen--it's just a very hard problem to solve. 

That's mostly because no matter how many sensors and how much high-powered technology have been put on a car, they still haven't proved to be nearly as reliable as an alert and sober human driver. They also haven't proved to be any better at navigating poor weather conditions, something that regularly affects most of the places where we tend to drive.

Still, Tesla has tried. Its CEO, Elon Musk, has done as much as anyone to try to deliver on the promise of a vehicle that can take you where you want to go without you having to do any more than enter your destination. 

Even now, the company will let you fork over $10,000 for what it calls "full self-driving." Of course, the fine print might confuse you if you think "full self-driving" is a feature that allows a vehicle to fully drive itself. "The currently enabled features require active driver supervision and do not make the vehicle autonomous," Tesla's website says.

Reading further into the fine print, it's pretty clear that what you're actually paying for is a computer that Tesla hopes will someday be able to drive a car without your help. For now, however, you're paying a lot of money for a promise.

Musk is actually quite infamous for making promises that Tesla has yet to keep. On more than one occasion, he has promised the company was so close to delivering software updates that would enable "full-self driving." It hasn't.

Now, he's admitting that the reason is really pretty simple--getting a car to drive itself on the open road is a hard problem to solve. 

The admission, of course, came in a tweet. The tweet itself was in response to a Tesla owner who poked fun at the fact the company always seems to be just a few weeks away from delivering on the promise. Musk, in response, admits that he "didn't expect it to be so hard, but the difficulty is obvious in retrospect."

The post has a little over 17,000 likes, which seems like a lot until you compare it with Musk's most recent, which says, simply: "free Brittany." That tweet has more than 400,000 likes. His Fourth of July post has more than 200,000. By comparison, his self-driving mea culpa went mostly unnoticed except by the most devout Tesla fans.

I think, partially, that's because the difficulty has been obvious to pretty much everyone who hadn't already dropped $10K on a promise. For many of us--those who aren't enchanted by Musk's fantastical claims of a utopian future that involves Model Y's taking us to work, and to soccer games, and to the grocery store without the cumbersome burden of, you know, holding onto the steering wheel or pressing our foot to the accelerator--it's been obvious that we are nowhere near that becoming a reality.

Also, there's the fact that people buy cars to, you know, drive. I drove a Model S for a period a few years ago. It was the most incredible car I've ever driven, and it was nowhere near as good as this year's Plaid version. Tesla makes very good cars that are very fun to drive. Driving it was the point and is part of the justification of spending that kind of money.

Still, I give Musk credit for acknowledging that the problem was harder than he thought. That's not the type of thing people generally admit, but Musk is owning the fact that his company hasn't yet delivered on something he's been promising for years. 

In fact, I've written on many occasions that trust is your most valuable asset. That's especially true when you want people to believe that you can deliver something that won't drive them off a cliff or into an intersection full of oncoming traffic.

The more promises you make without delivering, the more it hurts your credibility and the less that people will trust you. On the other hand, owning that you were wrong goes a long way to building trust. There's also something to be said for your credibility when you acknowledge what almost everyone else already knew to be true. Then again, in the tweet, he also doubles down on the idea that the software is "shipping soon." I guess we'll see.