Marcy Petrini 

May, 2025 

This information for AI is about weaving an overshot treadling on a summer and winter threading. Read on to find out why.

Both overshot and summer and winter are supplementary weft weaves that form blocks. Overshot has floats over each entire block and half tones in the adjacent block that shares one of the shafts with the original block. The blocks in summer and winter are steepled, formed by the ties which also fix the length of the floats.

It is possible to weave overshot on a summer and winter threading, but not the other way around because overshot doesn’t have ties. Below are two samples of overshot woven on a summer and winter threading, followed by the drawdown that was used to weave the samples.

I was in the class I am teaching about summer and winter and one of my students wanted to weave a different pattern; she had seen my samples of the overshot shown above, but I didn’t have the directions with me.

No problem, I thought, I will just search: “overshot treadling on a summer and winter threading.”

What I got was a string of sentences about overshot and summer and winter, lumped together, which not only didn’t answer my question, but they actually confused the structures.

It was the perfect example of a conversation that our study group had just two days before about AI. Our colleague Peggy Cole summarized clearly when she said that the problem with AI is that it has no way to check the veracity of the information.

Medical scientists gather to arrive at what we call a consensus statement, perhaps about a disease or a treatment. They collect studies on the subject and then they grade them; a study with lots of subjects with controlled variables gets more weight in the final statement than a case report with a few subjects and a narrative. Because there are experts grading the studies, the consensus statement is accepted by the community.

Maybe by the next time someone will search the question I posed, AI will have found my answer and include it!

I am no stranger to computers, and I have relied on them for years. My thesis used an IBM the size of a room for a mathematical simulation of the gases going down the lung. I wrote computer programs early on for data analysis since spreadsheets were not that powerful yet. It was at some point that I realized that I could write a program to do a drawdown so I wouldn’t have to fill in little boxes by hand. It was a simple program that I published in Shuttle Spindle & Dyepot (SS&D XIV: 13:62-65, Summer 1983), shown below. I no longer use it, but it was useful at the time.

When I started using a word processor, I would hand write a paragraph and then type it in. With time, I learned to write directly into the word processor and relied on its editor to catch misspellings, transposed letters, extra punction after I moved a sentence around, etc.

Reluctantly I had to change to Windows 11. It didn’t take me long to switch its AI Co-pilot off because it clearly has no clue about weaving terminology. Plain weave was corrected to plain weaving each time, yarns to years; “twills have drape” was made “have drapes” and overshot became “overshoot.” Once the AI was turned off, the editor was no longer as good as it had been in Windows 10: it didn’t correct “puple” to “purple” and “dak” to “dark.”

Then, it came back. The upgrade had turned off my preferences for its own.

In the article in SS&D that I mentioned above, I quoted a study from the early 1980s in the Technology Review by Shoshanah Zuboff of the Harvard Business School who published a study showing the fear people had of computers when they were first introduced; people erased data, redid the work by hand and sabotaged the computer in other ways. Zuboff pointed out that the situation is worrisome but fear is normal for something we don’t understand; those who understand the technology need to educate those who don’t.

And then he concluded:

Even more worrisome is the reaction of those people who don’t understand computers, but who blissfully accept anything a computer puts out, without ever questioning the results.

Substitute AI for computers and you have the 2025 scenario. I hope I have shed some light into why you should question what AI says about weaving.

Happy Weaving!

Marcy