The Art (& Tech) of the Message: from print to prompt

Several years ago, I curated a large exhibition on the evolution of the modern newspaper — the world’s first mass medium—as a graphic medium. The Art of the Message was based on rare, rag paper edition of Chicago Tribunes dating from shortly after the Great Fire in the early 1870s through WWII. The collection had been rescued from the Tribune dumpster by a sports memorabilia collector who wanted to cut up for vintage sports pages. This act of archival vandalism, made possible by a middle manager’s casual decision to clear out the company’s basement, ironically made it possible for me to work with the collection since it was no longer anywhere near pristine.

I filled up my car and hauled away as many of the large, heavy hard-bound volumes as I could, rescuing then from a storage locker near an expressway in suburban Chicago where they had been stacked and stashed. Then, thanks to a grant from the Illinois Arts Council and an assistant dean at Northwestern, I spent a summer clandestinely hidden away in an unused, former television news studio in the journalism building sorting through treasure: decades of history divvied up in daily installments.

Ads held a special fascination for me. Journalism typically focuses on reporting: breaking news, feature stories, investigations. This is what gets awards and why newsroom heroics became a Hollywood staple. Ads were the necessary evil providing the cash to keep reporters employed, the presses running and make publishers rich. Advertising was the devil, requiring an editorial “wall” to keep the truth-tellers in the newsroom safe and untainted from the spinmeisters on the ad-side of the business.

Yet it was advertising that led the way, constantly experimenting with how to make the most of this new, rapidly evolving medium. I found a front page from the 1870s—when lead type was set by hand onto printing plates—where six out of seven columns on the front page were ads. Typography was only way to stand out in the graphic crowd: “Fine Watches.” “Millinary!” And my favorite, “Cow Cow Cow”—all in fancy fonts and bold type.

The advertisers sparked an technological arms race among publishers to invest in better presses and printing technologies. Within a few decades, photos and elaborate engravings were standard, followed by color photos and lush rotogravure feature sections. Entire sections sprouted up to handle new markets. There were no cars in the 1870s, but the by the 1920s, cars were everywhere. The Tribune even started a comic strip, Gasoline Alley, in 1918 to tap into the craze. Remarkably, more than a century later, the strip is still going strong, although perhaps could use a refresh to “EV Alley.”

Ads also provided a kind of “carbon dating” time stamp. Cars models. Cigarette brands. Dish detergents. Movies. These were the products and services that defined daily life: the context against which the serious news of the day played out.

Today, algorithmically-targeted ads permeate the internet landscape. We literally no longer are all on the same page. Online, two people reading the same story from different devices will likely not see the same ads. The same reader may not see the same ads when a story is pulled up at different times. Pull up a story from the archives and the accompanying ads not only will be tailored to the individual reader, but also to the current moment, selling products and services that may not have been invented when the article was written.

••••••••••••••••••••••••••••••••••••

When the Tribune’s massive presses in the building’s sub-basements cranked up for the latest print run, it was said that you could feel it several floors up in the newsroom. Metal plates. Metal type. Special inks. Special paper. The Tribune was so vertically integrated, the company owned forests and mills in Canada to keep the news machine fed.

Well into the 1980s, newspaper production was still largely a mechanical affair. Rolls of film were developed in a lab, then photos selected from contact prints printed from strips of negatives to be enlarged. Illustrations were made with pen, ink and paint.

Designers, armed with X-acto knives, sliced print-outs of now digitally-generated type, which was then run through a machine that would coat the back of the paper with wax so it could affixed to a layout. Using photolithography, the layouts were transferred to printing plates.

Meanwhile, reporters were required to write to word length in order to fit within the limited “real estate” of a page or section. Likewise, ads were sized to fit the precise geometry of a layout.

By the 1990s, the mass adoption of personal computers and rise of the internet turned the the craft of production into a digital skill set: “cut and paste” without any need for sharp knives to actually cut or hot wax.

Photographers armed with digital cameras no longer were limited by the standard 20 or 36 “frames” in a roll of film. They could transmit photos from the field instantly. Video, too. Illustrators either scanned their work, converting it into bits and bytes, or created art digitally using specially developed computer programs. Stories now could run as long as needed, though confined to a screen-optimized, single column, endless scroll.

Once again, it was advertising pushing the boundaries of an evolving medium in ways both obvious (more interactivity) and subtle (A/B testing). Data, often riddled with bias, perhaps unintentional, but still scalable, became the unseen editor shaping content.

FROM PRINT TO PROMPT

The leap from hand-set lead type to digital layout took the better part of a century. The transition from photolithography to back-end computer code took decades. Generative AI, where bots call most of the content shots, is an overnight sensation.

In this Youtube video, a man who identifies himself only as being part of the “Codex Community,” gives a 20 minute demo on how to build a website campaign from scratch using Midjourney (art), ChatGPT (website architecture; copy) and Editor X (code-free website builder).

It is an astonishing feat. He types in a few, pithy prompts — “shoes, website, ux, ui” — and almost instantly Midjourney conjures up four options for graphics. He adds “Nike” as a prompt and “swish” logos appear on the sides of fantasy shoes so colorful and cutting edge that sneakerheads everywhere would line up to buy.

Next, Mr. Codex Community (Mr. CC) types a few prompts for ChatGPT and, voilà, copy is ready. Editor X is where it all comes together.

Mr. CC isn’t particularly concerned about the accuracy of the copy. Or, for that matter, a Nike copyright. This is, after all, a demo.

The next steps, not covered in the video, would be to link the new website to an AI-enhanced service auto-generating social media posts, and a data-torqued ad network to create a media campaign. Add payment and fulfillment services — presuming there is an actual product —and you could be in business. What used to take months of planning, or at least several weeks, can now come together in a matter hours for a fraction of the cost.

This is still early days in the Brave New World of AI, with much to be sorted out. But generative AI’s formidable, near-magical talent for making instant “art-in-the-style-of” illustrations and composing serviceable copy for everything from legal documents to college essays, press releases and new stories is a game-changer.

In a blink, marketplaces began popping up to sell “prompts” for generating content: the digital equivalent of sourdough starter.

•••••••••••••••••••••••••

Working on The Art of the Message, I thought a lot about whether the “art” was the photograph or illustration, the layout, the printing plate, or the final printed page, produced in editions that could top a million copies.

Now I wonder, is the “art” the prompt or the infinitely reproducible images and content the prompt generates?

THE “PROMPT” DIVIDE

In 1937, the Chicago Tribune printed a series of ads showing cutaways of the iconic Tribune Tower— then only 14 years old—as a vertical news factory. (The distinctive, limestone-clad Gothic highrise has since been converted into high-end condos.)

Near the top of the Tower were the cartoonists drawing Dick Tracy and Brenda Starr. At the very bottom, several floors below ground level, massive roles of newsprint rolled into the building on an underground rail system. In between, there was a newsroom, an engraving room, a color photo studio, a library and several floors for the ad sales staff.

Today, a handful of “Mr. CC’s” typing in “prompts” can do the work of hundreds, perhaps thousands of writers, designers, photographers, artists, researches and sales reps.

Generative AI will get better. It is getting better every day. And we will get better at using it.

The nature of work will change, which is nothing new. A century ago, you could make a good living setting hot metal type for printing. Forty years ago, X-Acto knife-wielding graphic designers were in demand.

How well one will be able to navigate the seismic shifts coming fast and furious depends on which side of the “prompt” you find yourself. “Intelligent automation” is just a new flavor of automation. Crafts that take years to master, whether building furniture or writing ad copy, will be undercut by much cheaper, “good enough” competition.

Personally, I hopeful that I am on the “prompt” side of the equation. My stock-in-trade has always been about peripheral vision: an ability to spot patterns and connections that have yet to come into general focus. This is the realm of the “adjacent possible,” a term coined in the 1990s by biologist Stuart Kauffman, who studied complexity.

As a generalist and a “silo-skipper,” almost everything is something about which I want to know more. I have written about everything from microbes to energy, chased wild horses/wolves/bears/coyotes, filmed salt sculptures in a salt mine, been on the business side of a tunnel boring machine, and ridden the rails with a team taking ultrasounds to check for metal fatigue. Unless there is a good reason to say no, I usually say Yes!

Generative AI is an extraordinary tool for looking at what has come before. But a computer database, no matter how vast, lacks imagination. In its current incarnation, ChatGPT can “hallucinate” and fabricate facts, but it cannot see outside the lines. Its forays into fantasy are algorithmic hiccups.

Imagination and creativity are about what could be. They weave together analysis, synthesis, curiosity and also humor. Humor rarely comes up in these sorts of discussions, but it is essential part of the equation. Humor is an evolutionary advantage, giving us away to explore the absurd: the logic of the illogical. We are also biologically wired to understand the non-linear, dynamic languages of shape, form, texture, color, light, touch, movement, scent and sound.

So.

I will bring the prompts. Generative AI can bring the databases. And who knows? Together, we might just reinvent the newspaper.