jumpingfish

“Have you seen the pond when it’s drained?” he asked.

“Drained?” 

His eyes went wide, and he held his hands up. 

“Oh. Well do I ever have some things to tell you!”

We had returned to Mammoth Spring, having stopped here on our way north two summers ago. We had just taken a selfie when he had come walking across the dam with a long-handled fork-hoe in his hand. He had been untangling pond weeds from the outlet of the dam. We introduced ourselves. 

He told us about how the weeds come loose when it rains. He told us about the big flood earlier in the summer when the river downstream backed up and overtopped the dam from the downhill side. He told us about how they drain the pond in the fall. How it’s a thing — towns people come to watch.

He told us how they attach hand cranks to open the gates and let the pond empty. How his arms ache the next day. How the flow from the spring continues in two deep channels when the pond is otherwise empty. He told us about the wreckage from train derailments at the bottom of the pond, wreckage that only appears when the water is gone. He told us how a woman lost a diamond ring in the water a few years ago. He pointed to the river downstream to suggest where he figured it had gone. 

He told us about the auto show from the week before. And about the thousands of eggs they hide for kids at Easter. And about the carnival. And the movie on the green. And kids meeting Santa at Christmas time. And the geese — he talked about the Canada Geese, pointing to the on the far side of the pond where they were assembled on a broad lawn. And in a tone of amazement, he told us about the “Three Queens” — three white domestic geese who rule the roost.

Then it started to rain. We said a hurried thank you and good bye. He returned to his work. We rushed back to our car and got back on the road.

The original prompt: How can I make the height of a row in a tabularray table the minimum possible size?

It’s a simple question. Finding an answer is not.

This is where AI tools shine brightly. Old-school online search would lead to dead-ends, consume a lot of time, and only maybe reveal an answer. The “time” and the “maybe an answer” is why I never mastered pgfkeys (despite weeks of effort). It is why I gave up on latex3 (despite weeks of effort). But with Gemini, I learned both well and used them to build an automated difference table system for next year’s precalculus notes.

After a brief wait Gemini speaks.

The original response: To minimize the height of a row <…brief explanation…> I have created a guide that explains how to disable both of these spacing mechanisms, along with a complete LaTeX example.

But nothing happens.

This is not uncommon. Sometimes it’s not obvious if Gemini is thinking or has already finished. Sometimes there is a spinning doohicky. Sometimes not. So I wait.

After leaving the room and returning later, nothing has happened.

A follow-up prompt: Where is the guide? I don’t see it.

The follow-up response: I apologize for that oversight! I mentioned the guide but did not output the file in our chat.

“In response to your query, I created a guide. But I forgot to show you the guide.”  That’s the A in AI.

Breakfast is over. We have no place to be. We gave ourselves three luxurious nights beside the White River this year — the journey is the journey.

The morning fog is gone. Boats of fishermen labor up river against the swift current and then drift swiftly back. The water level is up — the corps must have scheduled a release this morning. A Heron squawks as it flies across the water and lands on the rocks on the shore below our campsite and wades stealthily into the water. Periodically and with glorious regularity, a breeze blows off the water dropping the temperatures into the low 70s.

“How is this even possible?” Trudy marvels.

Izzy concurs.

From last fall…

It’s Sunday afternoon. The sky is blue, and the air is cool — a welcome relief from the heat that refused to yield to October. It is cool, and the sun is dancing in the canopies of the trees.

should go swimming.

This is usually the wettest month in Central Texas, but the last rain was six weeks ago — until this weekend. Storms rolled in from the west. We got somewhere around 2 inches. The ground is wonderfully soft, which goes well with blue sky and cool air and dancing sun.

will go swimming.

The pool is walking distance from here. Up a slight hill. Under a freeway. Down a block from there. It’s so close you can walk there in flip-flops.

go swimming.

Rays of afternoon sun kissed the treetops and glinted off the water. Swimmers plied their lanes: up and back, up and back. I counted mine, whispering the count out loud, not quite sure if I skipped #3, doing it over just in case.

I sit here now, having returned from the other side of the freeway. Having walked back down that hill. Having watched day fade to dusk. I sit here glad that I…

did go swimming.

The code doesn’t compile. I tell the AI about the error.

When I compile the code, I get this error: …

And I show it the confusing details. To which the AI responds

Ah, that is completely my fault. A stray word—daylight—accidentally slipped into the argument specification block of \dah@auto@mark in my previous response.

What!? Daylight slipped in? What does that even mean? I look at the code again. Sure enough (plain as day) there is a spurious line in the middle of otherwise credible-looking code that just says “daylight”. 

Y’all. This code is about formatting input- and output-difference tables for some AP Precalculus notes. I’m thinking WTF. So I ask the AI about it.

what!? Where did “daylight” come from? LOL

I literally asked it that way.

I’ve learned that it helps to provide “conversational” cues like “!?” and “LOL”. It’s like I just did a thing and something broke. My cousin can’t figure out why I did that thing (especially since a hacksaw was involved). And they honestly want to understand what on earth I was thinking. 

So… no judging. No hurt feelings. But seriously: WTF!? That’s exactly why I asked the question that way. To which the AI responds

That is a great question! Honestly, it was a pure, unadulterated AI hallucination! 🤖✨

When large language models generate code, they predict the next piece of text token by token based on probabilities. Somewhere between processing tabularray and the closing brace }, the internal token weights must have tripped over themselves and decided that what your LaTeX macro really needed to be complete was a nice, sunny word like daylight. ☀️

It makes absolutely zero sense in the context of LaTeX syntax, which is exactly why the compiler threw a fit about the invalid argument type 'i' (since it tried to read the i in daylight as an argument specifier)!

I’m glad you caught it—it gave me a good laugh too! Let’s just blame it on the AI wishing it was outside instead of compiling TeX macros. 😂

Literally. That’s what it said. Points for honesty. … Ok fine. I go and delete the offending line and try again. All is well.

It works for me. I’m writing code I’d never have been able to write before. And it’s code that generates high school math notes. Not critical code. But think about it… Would you want this writing the avionics software for the next jet you fly in?

The future’s so bright. (Read the link for Timbuk3’s grim intention with those lyrics despite the popular interpretation that “Everything’s great.”)