I recently got a look at the sweep of AI's progress through a fascinating book, "The MANIAC." It's a fictionalized oral history of the life of John von Neumann, a pioneer in mathematics, physics, computer science and more. He died young, in 1957, less than a year after the term "artificial intelligence" was coined, but the book carries his legacy through to today as a spark for DeepMind.
The AI company, founded in 2010 and now owned by Google, made a splash in 2016 when it dominated a match against the world's best player at Go, a game considered to be much harder for computers than, say, chess. DeepMind has since moved on to far more serious matters; two of its founders won the Nobel Prize in Chemistry in 2024 for a breakthrough in understanding how proteins fold, an issue that has massive implications for developing drugs to treat disease.
Much of the progress can be traced to the mind-boggling improvements in computer hardware — von Neumann's massive computer, known as the MANIAC, had all of five kilobytes of memory, while you may have 200 million times that amount just in your phone. The progress also obviously stems from the big brains, such as those at DeepMind, who've made AI their life's work. And the progress will continue at a furious pace, both because the computational capacity available to AI is doubling every three or four months and because AI has become perhaps the great intellectual challenge of our time.
But as much as I enjoy reading about intellectual titans like von Neumann and marvel at the sweep of history, the key question is: What are we humans doing with all this power that's being given to us?
Which brings me to a recent Harvard Business Review article that shows how companies really sank their teeth into generative AI last year and lays out where they can go from here in 2025.
The article, based on a survey of C-suite executives from 125 companies that the author has been conducting since 2012, opens with a bleak statistic that is all too familiar to those in the insurance industry: Just 37% of companies reported that efforts to improve data quality had succeeded.
But the article continues: "This year’s survey findings suggest that we are experiencing a once-in-a-generation transformational moment, akin to the founding of the internet in the 1990s.... 89% [report] that AI is expected to be the most transformational technology in a generation, up from 64% in last year’s survey."
The report describes six major changes in 2024, four of which I'll highlight here:
- 98% of organizations said they were increasing investments in AI and data, up from 82% last year.
- "94% report that... they are seeing quantifiable business results, which can be measured by metrics including increased customer acquisition and retention, improved customer satisfaction, and revenue and productivity improvements.... 18% report a high level of measurable business value, and 28% report rapidly increasing business value. Another 32% see a modest but rapidly growing level of business value based on these quantifiable metrics.... The source of this value is significant: 75% see it as coming from productivity gain and customer service improvement, notably through efficiencies resulting from the application of generative AI into traditional production processes."
- "Organizational transformation due to AI is seen as steady, but gradual. Most organizations characterize their AI efforts to be at an early stage, with 76% in experimentation, testing, and limited production. But... 24% reported implementing AI in production at scale this year, up from just 4.9% last year.... 91% of organizations report the greatest barriers to business transformation are due to cultural factors, not technology."
- While many companies are now appointing chief AI officers, "51% report that the AI and data leadership roles are not well understood within their organizations," and turnover is high. On the flip side, 36% of AI and data leaders "now report to the most senior business leadership of their organization," and the "AI and data leadership roles are now focused on business innovation, business growth, and business transformation efforts," not just on the technology.
A one-year snapshot isn't nearly as impressive as the 80-year sweep in "The MANIAC," but the HBR article still describes an awful lot of progress for one year, and 2025 is just getting going.
Cheers,
Paul
P.S. Von Neumann is a fascinating character, with an intellect that was intimidating, even scary. When he was three years old, growing up in Budapest, his parents would entertain guests by having them pick a page in the phone book. They'd show it to Johnny for a minute, then ask him anything. Who has this phone number? What is the phone number associated with this address? What is the address of this person? The question didn't matter. Little Johnny had the page memorized. "The MANIAC" describes a situation where a prominent mathematician told a class of students at maybe the world's preeminent technical high school that he wanted them to think about an unsolved problem that the mathematician had been wrestling with for years. As the students started to talk, 14-year-old von Neumann went silent. After about three minutes, he raised his hand, went to the board, and wrote down the answer.
Von Neumann gave us game theory, was one of the founders of operations research (which is how UPS and Fedex come close to optimizing how they handle all the packages they accept and deliver every day), laid out the mathematical foundations for quantum mechanics, figured out how in the 1940s computers could go beyond being hard-wired for every problem and could use "stored programs" (what we now call software) and much, much more. When he was dying of cancer, likely due to radiation he absorbed while working on the Manhattan Project (where he developed the implosion model used in the second atomic bomb), the Department of Defense surrounded his hospital room with security, lest he blurt out any of the numerous secrets he knew during his frequent bouts of delirium.
He also, according to "The MANIAC" and other books I've read about him, had no particular moral compass, unlike Robert Oppenheimer, the leader of the Manhattan Project. Von Neumann was just fascinated by the problem. He was the model for the main Peter Sellers character in "Dr. Strangelove."
Von Neumann was the prime example of the two sides of technology, the potential for almost magical progress and the simultaneous prospect of unintentional, dangerous consequences.