Let's Prove to Your Laptop That It's a Bad Laptop

(illustration)

Give it a good smack on the screen. Yell at it. You can even grab it by the screen and slam it against the table. That’ll teach the bastard. Try to unleash your anger properly — preferably on the latest MacBook or a Carbon X1, something expensive, so that you vent even more rage, and…

You’ll get absolutely nothing in return.

The theory of venting anger was stillborn. It worked well for preparing a boxer for a match and similar situations, but it absolutely does not work when it comes to calming someone down.

As far as I can tell, anyone with even a little observational ability can conclude that the opposite is actually true. Let’s run a thought experiment.

Sit down and try to analyze one of your latest screw-ups. Maybe you lost a client, crashed a server, or were late to class. Whatever it was. But instead of trying to justify yourself, do the opposite. Find the reason why, next time, in a similar situation, you’ll come out on top — or even prevent the situation entirely.

You can call the client back. You can back up the server, set up redundancy, or switch to different software. And you can leave for class ten minutes earlier.

See? You don’t need to find a reason why you were wrong in this situation. You’re already wrong. You need to find a reason why you’ll be right in the future.

I’ve heard countless stories of people who “reflected” because their HR forced them to. They sat there, telling a sad story about how Bill, Vasya, or Ahmed was to blame for their misfortune this time, and all that. But no one ever sat down and thought about how to be right in the future.

So: you can be right, you can achieve practically anything, and you can control objects in the physical universe. And if yesterday you couldn’t prevent a tree from falling on your car, then — come on — you’re an engineer. Figure out how to solve this problem. You were trained for this. You don’t need to run around the car, hitting it with a stick and yelling at it:

“You wretched piece of junk! How dare you get hit by a tree?”

Deep down, we all understand that inanimate objects can’t respond to us. No matter how much you yell at the remains of your car, you won’t get anything out of it. A computer won’t work better if you shout at it. You can yell at a child, and the child will only feel worse. You can yell at your wife, and she’ll hold a grudge against you. Yelling might help you if you’re a police dog. But yelling at inanimate objects, just like trying to talk to them, leads nowhere.

Now, carefully reread the last sentence and think about AI.

If AI is an inanimate object, then why do you expect that you can express emotions toward “it,” yell at “it,” or try to explain to “it” that “it” is wrong? A good ML engineer can always find an error in the input data. A bad one will ramble on about how AI has consciousness and how “it” just felt like it. But the last time I screamed at a server with all my might, it didn’t react to me at all.

If something is wrong with the output of your ChatGPT, Claude, or Grok, then you have only yourself to blame. You’re the one making this thing work. It’s under your control.