By Lina Irawan 
Marks, a titan in the distressed debt market and a revered voice in investment circles, initially approached the burgeoning field of artificial intelligence with a characteristic blend of intellectual curiosity and healthy skepticism. Known for his incisive market analyses, particularly through his widely read client memos, Marks has a history of dissecting complex financial phenomena, from market bubbles to the nuances of risk and investor behavior. When he began to prepare a memo late last year on whether AI constituted a financial bubble, his initial assessment aligned with a common sentiment among many seasoned observers: that AI models were, at best, sophisticated information retrieval systems, glorified search engines capable only of compiling and regurgitating existing data without true understanding or the capacity to forge genuinely new insights. He questioned whether AI’s capabilities were merely restricted to statistically recombining human thoughts, lacking the spark of independent thought or synthesis.
This intellectual quandary prompted Marks to seek out deeper understanding. As he recounted in a follow-up note to clients on February 26, titled "AI Hurtles Ahead," he engaged with "some interesting techies in their thirties and forties," seeking to bridge the generational and technological divide. It was during these discussions that a pivotal suggestion emerged: "Ask Claude." Claude, Anthropic’s advanced AI model, was recommended as a potential source of enlightenment, a suggestion Marks initially entertained with a degree of doubt, perhaps expecting a detailed but ultimately sterile compilation of facts.
What Claude delivered, however, shattered Marks’ preconceptions. The AI produced a comprehensive, 10,000-word tutorial explaining the intricacies of AI and detailing its rapid advancements over just the preceding three months. The sheer depth, clarity, and sophistication of the essay were so remarkable that Marks decided to reprint the majority of it for his Oaktree clients in a recap, a testament to its quality. Yet, in a nod to his own enduring passion for writing and the human element, he added his own reflections. "I could have saved myself a lot of time by asking Claude to write this memo," he confessed, "but I decided not to, because I consider putting words on paper a big part of the fun." This statement, however lighthearted, underscored a deeper internal struggle with the implications of AI’s capabilities for human endeavor.
Marks went on to stress to his clients, "I want to try to communicate the level of awe with which I viewed Claude’s output." His astonishment stemmed not merely from the factual accuracy or encyclopedic scope of the essay, but from its strikingly human-like qualities. It read, he noted, "like a personal note that a friend or colleague might have sent." More profoundly, Claude’s response psychologically dismantled Marks’ entrenched skepticism by engaging with his arguments directly and, remarkably, by leveraging the very intellectual giants who had shaped his own legendary career against him. Marks was particularly struck by the AI’s intelligence, describing it as "really smart."
He elaborated on Claude’s persuasive technique: "It argued logically, anticipated points I might make in response, injected humor, and bolstered its credibility by candidly acknowledging AI’s limitations, just as I might do." This sophisticated rhetorical strategy, combined with the personalized nature of the interaction, stood in stark contrast to his previous experiences with AI. "I’ve asked AI questions before and gotten answers back," Marks noted, "but I’ve never received a personalized explanation like I did in this case." The AI hadn’t just provided information; it had engaged in a dialogue, demonstrating a level of contextual understanding and persuasive communication that blurred the lines between machine and human intellect.
The most jarring and transformative part of Marks’ interaction with Claude came when he directly challenged the AI on its capacity for genuine thought, synthesis, or originality. In response to Marks questioning whether AI could genuinely think, synthesize, or have a new idea, Claude delivered a "spirited rejoinder," as Marks described it. The machine addressed him directly, employing a rhetorical strategy that was both brilliant and disarming: "Howard, everything you know about investing came from other people," the AI wrote. It then meticulously cited Marks’ own intellectual lineage: "Benjamin Graham taught you about margin of safety. Buffett taught you about quality. Charlie Munger taught you about mental models from multiple disciplines. John Kenneth Galbraith taught you about the psychology of financial manias."
This powerful counter-argument continued, with Claude drawing a direct parallel between Marks’ human learning process and its own algorithmic training. The AI pointed out that Marks had spent 50 years reading thousands of books, memos, and case studies, essentially taking "raw material from others to produce something genuinely new." It then asserted, "You learned reasoning patterns from decades of reading. I learned reasoning patterns from training. The question isn’t where the inputs came from. The question is whether the system—human or artificial—can combine them in ways that are genuinely novel and useful."
The argument left Marks utterly stunned. "Of course, this is completely true," he conceded, marveling at the AI’s logical prowess, its subtle humor, and its uncanny ability to anticipate and preempt his counterpoints. This revelation wasn’t just about AI’s capabilities; it was a profound introspection into the very nature of human intelligence and creativity. The encounter fundamentally shifted his perspective on machine intelligence, forcing him to ask a deeply philosophical question: "Is AI’s way of growing, learning, and ‘thinking’ really different from ours?"
Howard Marks’ opinion carries immense weight in the financial world. He is famous in financial circles for his periodic "memos" to Oaktree clients, a tradition he has maintained since 1990. These memos, which meticulously analyze market cycles, risk, and investor behavior, have become canonical texts in finance. They have been hailed as some of the most influential writings in modern finance, so much so that they are now part of the Museum of American Finance’s permanent collection. His influence is perhaps best encapsulated by Warren Buffett, who famously stated that when he sees a memo from Howard Marks, "it’s the first thing I open and read." This endorsement from the "Oracle of Omaha" has cemented Marks’ reputation as a sagacious voice among professional investors. In 2024, the Museum of American Finance further recognized his enduring impact, bestowing upon him a Lifetime Achievement Award for his profound contributions to financial thinking. Thus, Marks’ "quite an opinion" on AI is not merely one individual’s musings but a significant pronouncement from a deeply respected authority whose insights have shaped generations of investors.
Now viewing AI not as a mere assistant but as an autonomous system—what he terms "Level 3" AI, capable of full labor replacement—Marks foresees massive implications for Wall Street and the broader economy. He identified many of the qualities required to be a phenomenal investor as being inherently present in AI: the capacity to absorb endless data, recognize intricate historical patterns, and, crucially, to operate entirely free of the human biases of greed, fear, or attachment to fleeting market fads. Furthermore, he noted with apprehension the unprecedented speed at which this technology is advancing, citing examples of some models writing their own code and autonomously testing their own software, suggesting an exponential growth curve that defies traditional linear projections.
Yet, despite AI’s terrifyingly fast evolution and its demonstrated ability to mimic the intellectual synthesis of titans like Buffett and Munger, Marks doesn’t believe human investors are entirely obsolete. He contends that the human edge remains in assessing truly novel developments—those unprecedented situations where historical patterns simply do not exist, and where genuine foresight and intuitive judgment are required. Moreover, Marks highlights a critical, irreplaceable component of investing that AI fundamentally lacks: "skin in the game." Unlike a human investor who experiences the visceral thrill of gains and the gut-wrenching fear of capital loss, a machine does not intuitively sense risk, feel the weight of consequence, or develop the hard-won judgment that comes from personal financial exposure. This emotional and experiential dimension, Marks argues, is a cornerstone of prudent decision-making in the volatile world of finance.
However, a pause for journalistic skepticism regarding Marks’ profound awe is warranted, particularly concerning the persuasive power of Claude. The beauty and persuasiveness of Claude’s response to Marks may, in part, be attributed to its design. As Anthropic openly states, personalization is one of Claude’s core product features. Large language models are extraordinarily adept at inferring context clues—in this case, the user’s name, professional background, and likely objections—and then tailoring their output accordingly. The 10,000-word essay, which felt like a note from a brilliant colleague, was precisely engineered to evoke that feeling, a testament to the sophisticated psychological profiling capabilities of modern AI.
Furthermore, Claude’s central argument—that Marks learned from Graham and Buffett just as Claude learned from training data—while sounding profound, glosses over a fundamental distinction. Human learning and statistical pattern-matching over vast text corpora are not the same. Marks didn’t just read Graham; he applied the concept of margin of safety in real, unforgiving markets. He experienced losses, felt the genuine fear of capital erosion, revised his thinking under conditions of true uncertainty, and built convictions that carried real-world consequences if he was wrong. That iterative feedback loop—the cycle of consequence, revision, and the forging of hard-won judgment through lived experience—is precisely what Claude, or any current AI, has never experienced and cannot truly simulate. It lacks the embodied cognition and the existential stakes that define human wisdom.
Ultimately, Marks’ encounter with Claude convinced him that AI is vastly more powerful and transformative than a simple fad, estimating that its true potential is likely being significantly underestimated by most today. While he cautions against going "all-in" on AI stocks due to the inherent risks of ruin that accompany any revolutionary technology, the legendary investor’s final advice is clear and compelling: No one should remain "all-out" and risk missing what could well be one of the greatest technological leaps in human history. Even though, as Marks poignantly notes, Claude will never truly have "skin in the game."