When Katrina Manson’s Project Maven: A Marine Colonel, His Team and the Dawn of AI Warfare came out, it immediately landed on my next-read short list, my very short list, for two reasons.
First, I had followed Project Maven since Deputy Secretary of Defense Robert Work established it in 2017 as the “Algorithmic Warfare Cross-Functional Team.” At the time, the term “algorithmic warfare” was obscure and opaque, but at its core, consequential to our competitive edge in the future of warfare. Work’s vision extended well beyond applying artificial intelligence to intelligence analysis. Under his Third Offset Strategy, Maven sought to preserve America’s military advantage through AI-enabled decision making, human-machine teaming, and ultimately, as Work described it, institutionalizing autonomy across the force. The AWCFT (Project Maven) first mission was straightforward in concept but immense in execution: help analysts make sense of the overwhelming volumes of full-motion video collected in Iraq, Afghanistan, and elsewhere by using computer vision to identify, classify, and alert on objects of interest. Success here would establish principles applicable across command and control, logistics, maintenance, healthcare, administration, and virtually every military function. Maven was the first critical step.
Until now, relatively little had been publicly disclosed about Maven, and this book promised a rare inside view of the mission-driven team that dared to develop and prove these capabilities in the ultimate test of combat. Rather than relying exclusively on the methodical pentagon acquisition practice of exhaustive developmental testing removed from operational forces, Maven embraced a high risk and high payoff concept of continuous refinement alongside deployed users.
My second reason was more personal. I served with the book’s principal figure, Marine Colonel Drew Cukor, although I left active service before the Maven chapter of his career. I knew Drew as a prescient strategic thinker and a disciplined, if unorthodox, mission-focused zealot. Knowing his character and remarkable intellect, I was enthused to read the story of Project Maven from the inside and to see how his leadership of a small group of bureaucratic insurgents, operating deep within the Pentagon, transformed an ambitious vision into battlefield capability.
Project Maven did not disappoint. Every story has a beginning and Project Maven is a raw and remarkable account of what I expect will likely be judged as one of the pivotal moments in the Pentagon’s transition to an era of AI- enabled decision making.
Project Maven is Katrina Manson’s first book. An award-winning journalist specializing in national security, defense, intelligence, and emerging technologies, she previously covered the Pentagon and U.S. foreign policy with a particular focus on defense innovation, artificial intelligence, and military modernization. Her solid reporting credentials, access to key figures in government and industry, and understanding of the subject make her particularly well suited to tell Maven’s story.
Manson also deserves credit for earning the trust of the notoriously private Drew Cukor, whose perspective as Maven’s director is central to the narrative. Drawing on extensive interviews with the program’s principal architects and participants, together with years of Pentagon reporting, she reconstructs the origins, struggles, and significance of one of the military’s most consequential innovation efforts.
In Project Maven, readers are offered a front-row view of how a remarkably small, aggressive team, with top cover from Pentagon’s senior leaders, pursued an ambitious vision by recruiting talent, secured resources, built relationships with Congress, industry, combatant commanders, and deployed forces, and rapidly fielding bespoke capabilities at wartime speed.
Unlike many books on emerging technology, Project Maven discusses AI without becoming mired in technical detail. Instead, Manson focuses on the far more interesting challenge of organizing people, attracting industry partners, securing resources, and rapidly fielding software in combat through iterative development. From Somalia to Afghanistan and Iraq, readers watch both the technology and the confidence of deployed forces mature together. Remarkably, just eight months after Maven’s establishment, its first computer vision capability was operating in combat. Every improvement was shaped by operational feedback rather than laboratory assumptions. As bureaucracy tend to develop institutional “antibodies” to change, Maven succeeded not because resistance disappeared, but because mission focused leadership, and operational necessity consistently overcame it. Combat, not stateside testing, remained the ultimate measure of success.
But the most important story Manson tells is not about artificial intelligence. It is about why transforming the Pentagon is so extraordinarily difficult and why Maven succeeded where so many other initiatives failed.
