This piece originally appeared on Forbes.com on November 18, 2023
It seems as if every day, there are new predictions in the media about how moving forward into the future with Artificial Intelligence [AI] will change the present world of work that we know so well. Some articles see AI as a boon, removing jobs unfit for human employment, and speeding-up knowledge work, as well; while other articles worry about what will be left of human employment when AI is fully integrated into our work lives. As with all predictions of an unknown future, there is little guidance about which is more likely to occur, leading to AI becoming one of the most contentious issues of our times. Will it change our world, and in what way? Will it be destructive; displacing jobs indiscriminately? Will it render expertise a commodity, and diminish experts as it does so? Will it make us more, or less, intelligent? Will it concentrate special knowledge in the hands of large, faceless, organizations? Will it destroy teamwork? These, and many more, questions have appeared in popular and technical media to an extent that influencers and politicians, many of whom could not distinguish artificial from authentic intelligence, are proposing restraints and constraints on AI before we even have sufficient data to predict which of these outcomes are more likely to occur. Recently, a project which has received considerable popular attention, yet which deals with the ancient past, may provide a glimpse of some of the eventual ways that AI can effect our work; and the answers glimpsed are big!
Our quest for a better understanding of how AI will affect work in the future, begins, not with AI itself, but with a set of Roman scrolls, severely damaged in the eruption of Mount Versuvius in AD 79. The scrolls, now referred to as the Herculaneum Papri, were uncovered by a local farmer in AD 1750; and while many were destroyed in the discovery, about 600 remain, which, if readable, would represent “more than double the corpus of literature that we have from antiquity.” In addition, the possibilities of even more scrolls that are still unburied are high, and their decoding could literally change entire fields of ancient studies. It is possible that these scrolls once belonged to Lucius Calpurnius Piso Caesoninus, the father-in-law of Julius Caesar, but, no matter who their owner was, they had been so damaged, over the years, that they have been essentially unrollable, and so unreadable.
Enter Dr. Brent Seales, a computer scientist at the University of Kentucky, who has been working is this area of research for more than twenty years, beginning in 2002 with efforts to read an unrollable version of Beowulf in the British Library. In 2015, Dr. Seales was able to employ computer vision to virtually "open" similar scrolls from the Dead Sea region, but the writing on the Herculaneum Papri used a carbon-based ink that had no X-ray contrast against the underlying carbon-based scroll papyrus, making Dr. Seales’ previous successful approach unfit for them. As a result, the desired outcome was tantalizingly close, but still not accessible. In 2019, experimentation with three-dimensional CT scans, using a particle accelerator to pick-up surface patterns from writing on the papyri, finally resulted in a full virtual image of the scrolls to be produced.
While all this was going on in Dr. Seale’s laboratory, Nat Friedman, who was, at the time, the CEO of GitHub, an open-source software platform, had, independently of the University of Kentucky program, been turned-on to the ancient world by his reading of 24 Hours in Ancient Rome, “a book that he admits is pitched at eighth graders. I very irresponsibly stayed up late finishing this book…. and I had early morning meetings. But I just loved it,” Friedman has recalled. When Friedman learned of the project that Dr. Seales was working on, he jumped in, along with Daniel Gross, another Silicon Valley investor/entrepreneur, and several other donors, launching the Vesuvius Challenge, a crowd-funding effort, offering a series of cash prizes for accelerating the reading of the scrolls. In late March, 2023, driven by Twitter (now X) announcements, they raised the prize pot to considerably more than one million dollars, and, in the process, gave the Challenge enormous public visibility.
Faster than many thought possible, the Vesuvius Challenge generated results, and in October 2023, @natfriedman announced, on X, that a twenty-one year old, computer science student, at the University of Nebraska-Lincoln, Luke Farritor, guided by tweets from other contestants who were sharing their experiences on-line, "became the first person in two millennia to see an entire word from within an unopened scroll." Using data provided on the Vesuvius Challenge website, and a machine-learning model that he had developed, Farritor was able to decipher ten legible letters, within a four centimeter square, on the scroll, earning him the $40,000 prize for first ten readable words. As he worked, Farritor, in turn, posted his progress on X, allowing twenty-six year old, Youssef Nader, a computer science Ph.D. student at Freie Universitaet, in Berlin, to eventually win $11,500 for second place, but which also allowed him to make progress towards the ultimate $700,000 grand prize for “at least 4 separate passages of continuous and plausible text from the scrolls, each at least 140 characters long,” Professor Seales, who at one time had observed that “For me, reading words from within the Herculaneum scrolls is like stepping onto the moon,” now believed that the reading of the entire first scroll could well be concluded by the end of this year , thanks the application of AI and bright, young talent, and their experiments.
All of this makes for an exciting story, to be sure, but how does it help us understand how AI could affect the nature of work in the future? More than a few insights can be gleaned from this experience:
· Insight 1: AI can open-up entirely new fields of old and new work: Small as this project is, it is, nonetheless, a real example of how AI can open-up whole new means of exploration, be it in machine learning, or in ancient literature. Neither advance was necessarily expected, but neither is insignificant. And, AI does not have to necessarily be labor-displacing. Federica Nicolardi, an assistant professor at the University of Naples, and a participant in the project, sees big changes in her field as a result of how this project worked, and what it is finding: “This will change papyrology in general… It opens up a new part of the discipline.”
· Insight 2: The project’s success is the product of a community of efforts, Coordinated by.... [correction! because there is not conventional coordination, by anyone, in actuality; this is really a multi-centric community. So, we restart the sentence] Revolving around, the basic exploratory work of Professor Seales, what we see here is an ecosystem built upon the acceptance of informal multi-institutional partnerships, none of which could have succeeded without the rest. A professor in Kentucky, working on Beowulf in London; young computer science students in Nebraska, Berlin and elsewhere; the serendipitous appearance of Silicon Valley investors; crowd-sourcing on Twitter; this was not a carefully planned, pre-ordained, solution dictated by a formal project strategy; and no sane strategist would have plotted it out in just this fashion. Yet, it has worked successfully because the right people, with the right skills, have converged, at the right time, but for different reasons, on the same target.
· Insight 3: Big change often occurs in unexpected places.. on the edge of an established industry. It is no exaggeration to suggest that the Vesuvius Challenge is far from the center of either the academic or AI communities that will be the primary beneficiaries of any success that this project produces. But, it is also not exaggerating to suggest that the literature unearthed, the algorithms developed, the careers accelerated, and even the role of crowd-funding for academic projects, created by this project could well profoundly shape the future of several fields in the future.
· Insight 4: Leadership generosity in sharing opportunities can accelerate the accomplishment of work that has previously resisted prior efforts. This project has accelerated beyond the wildest hopes of the participants, largely because of the openness and sharing of ideas, information and opportunities within the project community. Professor Seales’ willingness to allow so many others into the project, especially at its most visible phases, is a testament to the rewards associated with leadership generosity. So, too, is the participation of the investors who launched the Vesuvius Challenge, motivated mostly by an opportunity to add to our knowledge of the ancient world, which has been an act of great generosity. These are good illustrations of how leadership generosity will become a powerful advantage in attracting ecosystem partners and external funding as we all go into the future and need to explore temporary partnerships of so many unfamiliar expertise domains.
· Insight 5: We are witnessing, again, the benefits to be derived from Open Innovation. Open innovation is not new, but like any non-traditional idea, it can benefit by a reminder of how powerful it can be. Prize-offering, as part of open-innovation is also not a new phenomenon. The challenge of determining longitude at sea was solved by eighteenth century, clockmaker, John Harrison, who won a £23,065 Longitude prize; and, in 1956, the winning design of Danish architect Jørn Utzon for the magnificent Sydney Opera House was chosen from 233 entries from 32 different countries. The acceleration of the Vesuvius Challenge was due, almost entirely, to the motivation of outsiders who were able to bring their expertise, willingness to experiment and time to the project. The machine learning techniques that the contestants used also weren’t particularly new, but bringing an "open-source mindset to an academic project" is what enabled such rapid achievements, according to JP Prosma, a spokesperson for the contest.
· Insight 6: Experimentation is the best way forward in the unknown. There was no project plan for this project. From beginning to end, was all about experimentation. Professor Seales’ early work with the Beowulf text, continuing efforts to “open” other closed scrolls, the machine learning models that different Challenge participants tried, and the sharing of results within the community, all speak to the value of continued experimentation throughout the project’s life.
· Insight 7: the introduction of AI might mean the end of linearity and sequentiality in discovery: In The Code Breaker, Nobel Laureate Jennifer Doudna, questioned our reliance on a traditional notion of linearity in scientific research. The Vesuvius project is an excellent example of a non-linear invention journey. In this specific case, AI, and Open Innovation, are generating surprises that can accelerate project performance, and from many different links in what used to be considered a necessarily linear, sequential, phenomenon. As a result, we are witnessing a long-accepted model of Science, which is often painstakingly slow-moving and linear in style, insight building upon insight, being upended by this project. In the words of one observer “in a few months, a handful of people who had never worked in papyrology moved the field forward twenty years.”
· Insight 8: Maybe the individual inventor is back?… This project team was about unconnected individuals, possibly inventors, each, autonomously connecting by informal means (Twitter, National Public Radio, crowd-sourcing and crowd-funding) and forming a temporary ecosystem of partners around a specific target; they were a virtuoso team, in fact, without anyone incurring the fixed costs of hiring. And, this ecosystem was, and is, inspired not so much by what they are being paid, or can win, but about what success will do for them in the future: it is a classic “We win, I win” arrangement. In addition, the allure of knowing more, is so powerful, that in almost every instance, the prize winners reported that they were going to use their prize money to buy bigger computing capability to continue in the project. Luke Farritor, was quoted as saying that “winning isn’t everything. “I just want to read the scrolls. We all just want to read the scrolls.”
· Insight 9: this team will likely never work together again! Don’t be nostalgic about such virtuoso teams. This particular team came together, autonomously, to address a specific project, and will likely never work together again, nor should they. They were the right team, for the right time, and once this project is finished, they will no longer, necessarily, be that right team. This is very much the epitome of what leadership scholar Warren Bennis once called The Temporary Society, 1n 1968.
· Insight 10: Looking around, rather than ahead, is a good way to think about the future: This is the advice of Silicon Valley icon, and former Director, and Chief Scientist, of Xerox’s fabled Palo Alto Research Center, John Seely Brown III, who advised seekers of foresight that “the way to look forward is, paradoxically, not to look ahead, but to look around.” It’s good advice, because few of us can look into an unknown future with sufficient confidence that what we are predicting will be anything like that which will ultimately unfold. The Vesuvious Challenge story is just such a sideways look around; a story of a project, presently in motion, and succeeding, that offers an optimistic view of how AI can work to create along with a talented team of humans, organized in unconventional ways to create a positive outcome, and yields leadership lessons about how we can increase the likelihood of other similar projects succeeding in the future.