mike's website

writing, mostly about technology

<<<

Cheats Ruined The Game

I remember when I used to play video games. I played lots of games. I have an early video game memory. It was around the 1999-2000 timeframe, my dad had a Windows 98 PC in the living room. We had Warcraft 2 on that computer. I loved Warcraft 2. I used to set up a single-player game that was a little bit easy, and I would painstakingly beat the enemy back into a corner of the map – just enough so that they would still be alive, but so that I could wall them off. Then I would build build build; I spent hours just building all of the buildings. It was an RTS, but I didn’t care. I just wanted to build.

I recall when I first found out about cheat codes. It seemed too good to be true. I could become invulnerable, I could build entirely free of interference from the enemy. I could build powerful units for free and sweep the enemy from the very start, painting them into a little corner. Then, I could set to building.

It seems like using a cheat code is going to improve the game, right? Glittering prizes, there’s your pack of Paladins. But even seven-year-old I discovered very quickly that when I cheated of course it sucked the joy out of the game. It felt too easy. With single player Warcraft 2, knowing that I could cheat, just by typing, was a complete fun killer. It was a good lesson for seven years old.

Now, I’m grown up. I became a programmer. There’s lots that I like about programming. It’s solving a puzzle, it’s being creative with a design, it’s intentionally communicating with others. The feeling of programming is a flow state, a mental exercise, a kind of state of mind, not unlike the way I used to feel about games. Building.

And now, we have AI programming. The AI is pretty good at programming. I wanted to write a package in Rust that would parse out symbols from a dynamic library — to build FFI bindings from an already compiled library, with a program. This is a pretty difficult thing to do, and in fact there is no existing tool that I can find that is capable of it. (You have to compile the library with debug symbols — not common, which is part of the reason it hasn’t been done before I suspect.)

So I assigned Claude code to do this. Claude code is pretty good in my experience. Of course, I’m still in the drivers seat, right? Certainly that is how Claude Code is marketed. Well, this is a pretty difficult task. It takes a lot of effort. I ask Claude for a first pass, some ideas. I know these debug symbols exist; there’s a library to read them. Claude has heard of this library, it knows how to use it, it can come up with some ideas. Claude came up with some ideas. I know it saved me time – I will defer to Claude because it knows the library better than I do. Why not? I asked it to write the implementation too. Pretty soon, I find that the tool works pretty well. Version 1 is done. With my guidance?

And this kinda bothers me. I knew I could have done this myself. It would have taken me a couple months of consistent work, probably. I would have had to learn the ins and outs of debug symbols, object files, the libraries that read them, the tree to traverse, the datastructure that could represent functions and their parameters and return values. It would have made for a fantastic programming exercise — a great game. Claude beat the game in 3 days with some of my promoting. I cheated. I learned nothing. I could not write this package tomorrow without AI assistance, even though I have already written it.

I have a cheat code now, and the game is not fun any more. Except now the stakes are higher. It’s my whole career. Management knows we all have the cheat code — it’s expected. Soon, AI programming will be very good and it will be everywhere. Gittering prizes.

Of course, I’m worried about my salary. Will my skills still be in demand with all of the AI assistance? Not too sure. But I’m not really worried about making a living. I know that I will be able to live in this world. I’m an AI skeptic, but at the very least the world today, on the eve of the so-called AI transformation, is broadly rich by every objective standard. There will be food on my plate if the machines take my job.

So, I’m not upset that the machines (might) take my job away. The problem is that what I really wanted to do with my life was play games — without cheats. That was the world that I wanted. A world of meaning. It feels somewhat that cheats have ruined the game.


Woman Holding a Balance

People used to make clothes. You’d have a couple pieces of clothing, you’d have one nice suit for church on Sundays. And then, over the course of a generation or two they mechanized clothing. They made it so that everyone could have a Sunday suit. They made suits so cheap that they disappeared.

By all accounts, having more clothing was good. But, in the Industrial Revolution people started to notice that the meaning of their lives was seemingly being taken away. Lots of literature was produced dealing with this, trying to come to terms with it. Marx, alienation; Nietzsche, nihilism; Heidegger, Gestell. They came to articulate it pretty well, remarkably well, all of those works that deal with the cultural and psychological fallout of industrialization did a great job of characterizing what was certainly a very diffuse and hard-to-articulate problem – it was this collective loss, this psychological loss. The nice Sunday suit not being quite as important any more, now that you could buy one off the rack. What is one’s place? What is meaning?

There is always a little worm in my mind about those works of art from the past. As clear as they are, as transcendant and accurate and true; the worm is that they can describe the problem, they can articulate it in so many beautiful and accurate ways. But they never can solve it. They could describe the feeling of a Sunday suit not meaning any more, the implication of it, the shape of the loss. But they could not bring back the meaning. It’s gone forever.

The march of technology is a story about the retreat of meaning. An accelerating march, and an accelerating retreat. Now, it seems today that meaning is in short supply. They said that back then, too, but an exponential looks the same no matter the scale you look.

With AI, and programming, I see it. I see the future and it does not look good. Well yeah sure, if you’re looking at points in the game, it looks utopic. It seems we have discovered yet another cheat code, a big technologic leap.

If the boosters are to be believed, this technology is going to change everything about our lives. The way we work; the way we create art; the way we communicate; how we work and play. How we relate and think of ourselves.

But the meaning, the thing that makes the game, that will retreat. I’ve never been more certain of anything. This retreat of meaning, it is accelerating. Even the most brilliant minds 100 years ago could not figure out how to stop it. They could articulate it but they could not stop it.

What the am I supposed to do about that?