Are Stendhal and Shakespeare ready for the world of AI? Mike Gioia says “yes”!

Bringing Stendhal to the 21st century

Mike Gioia wants to broaden the reach of poetry through digital media. That’s why he created a poetry film studio called Blank Verse Films (you can find them on YouTube here), where he experimented with new, ambitious ways to bring poetry to audiences by adapting it into short films. He also founded a generative AI company called Pickaxe.

Name sound familiar? It should. He is the son poet Dana Gioia, former National Endowment for the Arts chair and former California poet laureate. I’ve written about him countless times on the Book Haven, for example here and here and here. Type in the search box for more.

Mike and I have something important in common: both of us share a love of literature and the humanities. Moreover, we’ve both received Emergent Ventures grants from the Mercatus Center, the creation of Tyler Cowen. Mike is one of the most recently honored by the grant program.

I’m a big advocate of video as a mass education tool and way to reach a broader audience,” he says. “I pushed my to dad to film and publish all the poetry videos on his YouTube channel.” (You can watch them here.) “I’m focused on making super powerful tools like Large Language Models accessible to ordinary people through good, simple design and practical applications.”

From Mike:

I’ll win Book Haven readers’ trust with an uncontroversial opinion: reading Shakespeare is enjoyable and worthwhile. And now I’ll lose you entirely: we should read Shakespeare with AI. 

On set of a comedy pilot, watching the director’s monitor.

Recently, I’ve been reading with the assistance of Large Language Models (abbreviated to LLMs). These are AIs that can process and generate text. 

When I use LLMs to read, I’m still reading. But AI is on the sideline to contextualize anything I want, transform paragraphs into new shapes, or even offer an opinion. LLMs are models trained on the entirety of humanity’s literary output with the goal of “predicting the next token” in any sequence of text. In other words, they’re super intelligent auto-complete systems. It’s surprising how agile these models are with such a narrow mission. They can simplify complex ideas, modernize archaic expressions, and even transform subjective prose into objective statements. This versatility proves invaluable when reading texts that may be challenging due to historical or cultural contexts. 

Take the example of French realist novels. I’ve always loved The Red and the Black by Stendhal, but my less than impeccable knowledge of the Bourbon Restoration in French made me suspicious I was missing context. So I created an AI tool with the instructions: “Please explain any context necessary for a modern reader to better understand the following text from Stendhal’s The Red and the Black.” 

Here’s the opening of Stendhal’s novel: 

The little town of Verrières can pass for one of the prettiest in Franche-Comté. Its white houses with their pointed red-tiled roofs stretch along the slope of a hill, whose slightest undulations are marked by groups of vigorous chestnuts. The Doubs flows to within some hundred feet above its fortifications, which were built long ago by the Spaniards, and are now in ruins. Verrières is sheltered on the north by a high mountain which is one of the branches of the Jura. The jagged peaks of the Verra are covered with snow from the beginning of the October frosts. A torrent which rushes down from the mountains traverses Verrières before throwing itself into the Doubs, and supplies the motive power for a great number of saw mills. The industry is very simple, and secures a certain prosperity to the majority of the inhabitants who are more peasant than bourgeois. It is not, however, the wood saws which have enriched this little town. It is the manufacture of painted tiles, called Mulhouse tiles, that is responsible for that general affluence which has caused the façades of nearly all the houses in Verrières to be rebuilt since the fall of Napoleon.

And here’s the context my tool gave me: 

To better understand this passage from Stendhal’s The Red and the Black, it is important to consider both the historical context and Stendhal’s literary style.

Historical context:

Mike Gioia directing on set.

–– Franche-Comté: Franche-Comté is a historical region in eastern France, bordering Switzerland. It was an important strategic location between France and the Holy Roman Empire during the 17th and 18th centuries. The region was known for its natural beauty, which Stendhal describes vividly in this passage.

–– The fortifications built by the Spaniards: During the 16th century, Franche-Comté was controlled by Spain as part of their Habsburg territories. Spanish forces constructed numerous fortifications throughout the region, including at Verrières, as a way of establishing their dominance.

This context is several steps more specific and more useful than writing a comparable google query. I was able to get this information within seconds without much effort. 

Now, there’s an obvious objection that using AI to translate classic works degrades the action of reading. This perspective misconstrues the role of AI in reading. In most cases, AI should be seen as an augmentation for an activity rather than a replacement for it. AI reading co-pilots promise to drastically expand the readership of many older literary classics. The opportunity to grow audiences is especially exciting with Shakespeare. While Shakespeare perseveres to become a favorite of anyone who gives him a fair try, for a lot of readers the Elizabethan language is a barrier to entry. Phrased more bluntly, Shakespeare is hard to read for first time readers! When they get it, they love it. But they have to get it first.

I’ve always maintained that the message of poetry is universal. And I’ve done a lot of work to bring poetry to wider audiences. It’s with this same mission I sat down to build an AI-powered Shakespeare Translator on Pickaxe to help young readers enjoy the Bard. The tool allows readers to instantly translate any Shakespearean text into modern English. The tool is not rewriting Shakespeare. It’s offering a plain English explanation for any chunk of language that isn’t transparent to a reader. These are not attempts to supplant the original. They present a simple interpretation of the original passage that maintains the original message and themes, and allows readers to return to the original text with enhanced enjoyment. 

Mike encourages everyone and anyone to try it or use it in classrooms. You can try the Shakespeare translator tool on Pickaxe here. Let us know how it goes.

Postscript: “So how does it go?” I asked. Like greased lightening. Mike Gioia is already in The Guardian, as of a few days ago. From the article: “Those who hate AI are insecure’: inside Hollywood’s battle over artificial intelligence”:

Some recent entrants to the AI industry say that the current technology is being overhyped, and its likely impact, particularly on writers, has been exaggerated.

“When people tell me the studios are going to replace writers with AI, to me, that person has never tried to do anything really difficult with large language models,” said Mike Gioia, one of the executives of Pickaxe, a new Chat GPT-based platform for writers with a few hundred paying customers.

He called the idea that AI could produce full scripts “science fiction”.

“The worst-case scenario for writers is that the size of writers rooms is reduced,” he said. …

Writers have made AI central to their strike in part because “it’s a good story”, Gioia argued and partly because they are much less accustomed to being disrupted by technology than other industry workers.

“A lot of people in post-production have lived through multiple technological revolutions in their fields, but writers haven’t lived through a single one,” he said.

Read the whole thing in The Guardian here.

Tags: , ,

One Response to “Are Stendhal and Shakespeare ready for the world of AI? Mike Gioia says “yes”!”

  1. John Keyman Says:

    Mike Gioia’s assertion that Stendhal and Shakespeare are ready for the world of AI is certainly an interesting one. The works of Shakespeare have been used as training data for natural language processing models for many years now. His language, with its richness and complexity, has provided a robust corpus for machine learning algorithms to learn from. Similarly, Stendhal’s writing style, with its focus on psychological realism, has also been studied by researchers and used to improve AI models.

    However, it is important to note that while AI language models have been trained on the works of these authors, they are not a replacement for human creativity and interpretation. The nuances and subtleties of language, which make these authors so enduringly popular, are difficult for AI to fully replicate. While AI may be able to generate text that resembles Shakespeare or Stendhal, it is unlikely to match the depth and complexity of the original works.

    While Stendhal and Shakespeare are certainly relevant and useful for the world of AI, they are not entirely replaceable by it. The human creativity and interpretation that these authors represent will continue to be valued and appreciated for generations to come.