AI

There are 2 posts tagged AI (this is page 1 of 1).

AI is Not Inevitable

Last week I told my manager what I really think about AI and LLMs. I had shared my concerns before about certain uses we make of it, but this time was different. He asked, and it just came out. The built-in biases, the social harm, the energy usage, the mind boggling eagerness to use a tool that is set on making us obsolete.

For a bit afterwards the thought “holy shit did I say that out loud” was going on in my head. But not for long. I’m glad I said it.

Recently I was going over my notes from 2022 when this bandwagon started rolling. If I can say so myself, it struck me how prescient I was.

I’ve used LLMs and I’ve read a lot about them, from people who retreated from their use as well as people who continue thinking they are useful. I am listening. But I feel good that I still have the same qualms and concerns today that I had then.

The Earth is not an infinite resource. Tech cannot be trusted to be socially responsible. And no it’s not inevitable.

The only way to win a con is to not play.

That’s exactly how I feel about it. Thanks to Mike Monteiro for putting it into words. I’m in my fifties. I know how to write. I know how to design. I know how to code. I know the value of a human’s creative output. After four years of watching the AI thing unfold and trying to be measured, I still refute claims that it’s inevitable and I still don’t see why telling a minion to do something for me has any value.

I’m amazed at the lack of conscious questioning that people have versus this tech. People willingly jump onto the bandwagon, seemingly unaware or unconcerned. “But look what it can do!”

I just don’t agree with that. I don’t think AI is a good thing, and I’m not going to force myself to use it just to pretend it protects my employability.

People seem to think that the market will automagically decide for them while they can remain on the fence and quietly go along. That’s not how that works. If you have concerns or don’t agree with it, say so.

Automatic Code Degeneration

At first I didn’t understand why some software developers are excited about AI code generation. If the stated goal is to make them obsolete and take away their very livelihood, why are they all for it?

Take authors, people who write books for a living. Should they be excited that an LLM can write an entire novel faster and more automatically than they can? No way. The act of creation, of sharing their humanity and lived experience is central to authoring books. Without that careful crafting of the words, there is no outlet anymore.

So what’s the difference? Me, I’ve always seen writing code as similar to authoring words. Coding is writing. It’s an act of creation, of crafting abstract ideas into a consistent whole that other people find useful. It’s a manifestation of the fact that I was here and I produced a collection of words that mean something.

And thinking about this, I’ve come to the overly simplified theory that there are broadly two kinds of software developers: those who like the art of writing code, and those who see it as an obstacle. The former are in this profession for the love of designing and building and tweaking and solving puzzles until they get everything right. The latter are more interested in the result, in the fact that software can exist one moment when it didn’t before, and they have less regard for how it got made.

So given the choice, would I prefer to use software made by people who used every shortcut possible to get a product they can profit from? Or by people who love the craft of designing and coding and solving human problems?

The latter, hands down.