I feel a sense of existential dread. Not the old dread -- although that never went away -- a new one.

Many years ago, skilled tradespeople worked hard to build great things like chairs and bookshelves and bed frames and sinks. Human beings would devote years and years to perfecting their craft and they became profoundly good at what they did. They mastered woodworking and masonry; they worked with iron and steel and copper and stone.

Have you ever watched a skilled tradesperson ply their skills? Have you watched them scrape a piece of wood or blow a piece of glass without machinery or computers or guides? Have you watched them create something that can only be described as art using only a few hand tools?

I have a dresser that's made out of oak. Only oak. It has no nails. There are no screws. The entire piece is made of wood. The joints of the drawers are dovetailed perfectly. The dowels that secure the top in place are immaculate. It was my grandmother's for many decades. The piece is well over 100 years old. The only imperfections on it are where I did a poor job refinishing it twenty-something years ago.

The people who made that dresser excelled at their craft. Given the condition of the dresser, it has at least a couple hundred more years left in it.

Something happened along the way. What was once the sole purview of skilled tradespeople became subject to mass production. In and of itself, that's not a bad thing. However, instead of one or two people knowing all there was to know about making a dresser, there became a bunch of machines that would cut and trim and glue and nail things together. Those skilled tradespeople needed to know how to operate a machine, not how to build the whole thing from scratch.

It's understandable. I don't begrudge the growth.

I also understand that not everyone has the privilege of owning a dresser built before there were such things as World Wars. I understand that sometimes, you need a dresser right now and you don't have the resources to purchase something that took person-weeks of effort to build. I understand that sometimes you can't use solid oak; that the budget allows for fabricated materials made of glue and sawdust. I don't judge those situations at all.

I was having a conversation with someone earlier about creativity in the realm of artificial intelligence. They were someone experienced in visual art. They're an artist. Their skills and experiences and talents are unknowable to me. We spoke about how they've been resistant to adopting AI; they recently took a course that encouraged a survey of generative tooling that could create text and images and videos. I sat virtually alongside them as they went through their course and all I can say is, "wow." There's so much out there. There's so much to know and it keeps changing.

Read More

Documentation Is Not Dead in the Age of AI. It Matters More Than Ever.

LLMs (Large Language Models) and AI are very good at producing code quickly. Sometimes, AI can generate in minutes what would take an experienced developer days to write. It's similarly very common to use AI to "update this function to..." or "modify this data structure to..." or "optimize this code to..." and let the machine do the heavy lifting.

It's tempting. It's really tempting. In fact, not only is it faster and easier to let the computer worry about edge cases or obscure syntax, there's often pressure to get more done, faster. Some places of business mandate the use of AI. There's definitely an incentive to "do more with less." Would you rather pay a human for hours of effort to write a function that may not address edge cases and may not be optimally efficient, or would you rather have an LLM spend twenty seconds to put together something quickly? After all, LLMs are trained using samples of code that were already written, already checked for edge cases, already optimized, and already made available to the public; why not take advantage of those existing efforts and investments?

So, if an AI will be drafting code, who cares about documentation? Why would someone take even more time to explain their code, document their choices, work through the details and consequences of a block of code? If the incentive is to get more code written faster, if the consumer of those comments will most likely be an AI that can figure out how things already work on its own, why in the world would someone look for new and creative ways to make that process take even longer?

That line of thinking is understandable. It is also wrong.

Read More

I recently wrote about how LLM hallucinations are unfortunate, several failure models, and how to work around them. This is a rewrite of that article for a more technical audience. Same advice, same problems, just worded differently.

Long Processing

This one is my least favorite pattern. It sounds so reasonable and so plausible. When asked about it, the LLM's responses can sound positively rational and well-grounded.

Example

It often looks like this:

Read More