The other day I was putting together a talk for the ePRST conference, and I decided to try building the slides with AI. I was pleasantly surprised - the slides came out striking, and more importantly I didn’t have to spend time on the thing I hate most: hunting for images in the right format and color scheme, nudging and aligning blocks, fixing a layout that keeps falling apart. All at once that misery of design and formatting was behind me, and I did the main thing - I brought my idea to life, and without much effort or negative emotion. The process itself felt like working with a designer.

That reminded me of a conversation with a colleague a year ago. We were discussing how the era of MS Office as the main interface for working with information is, apparently, slowly going to fade. Documents will stay, but the need to fight the form by hand will go away. Back then it sounded nice, but theoretical.

And now I can say with confidence that preparing slides has stopped being a painful process for me and has instead become creative work in a pair. The core skills turned out to be the ability to state a thought clearly, to spot where the AI hallucinated, and to finish it to my own standards.

It looks like the set of formats itself will change next. Text, document, slide - these are still a legacy of office suites and the age of A4. Once AI starts choosing the medium to fit the meaning, a thought will no longer run into ready-made templates. In some cases it will stay text. In others it will become a short video. In others - an interactive audio-visual format. There’s a lot of talk now that AI will reinvent processes, products, business. The very formats for presenting meaning are, in my view, part of that same list.

And here an old anxiety surfaces - will we forget how to write, draw, lay out a page? For me that anxiety is like tears for Assembler. The arrival of high-level languages did not kill programmers’ thinking. It removed the need to speak to the computer on its own level and made possible systems the old approach could hardly have reached. With the form of documents it’s probably the same story.

In my experience, AI amplifies whatever you hand it. My choice is to keep the idea for myself and hand the form to the AI. Then AI doesn’t wean you off thinking; it lifts away part of the secondary, routine load. But if you hand the AI the idea itself, then of course there’s a risk that part of your own thinking will gradually leave along with it.