AI Is a Tool: Why Mastery, Not Adoption, Defines Success

The Great AI Tool Debate: Why Rich Hickey Is Wrong (and Right)

In October 2026, software visionary Rich Hickey, creator of Clojure, sparked a heated debate with his article "'AI' is Not a Tool." His argument: LLMs are automation, not tools, because they cannot be truly "wielded" by a human operator. The response from the developer community was swift and divided, but a closer look reveals that both sides are missing the crucial point—AI is absolutely a tool, but using it well demands a level of skill that most organizations and individuals have yet to develop.

Hickey's concern is rooted in a valid fear: that outsourcing cognitive work to AI will atrophy our abilities to think, learn, and create. But as a detailed rebuttal on yyhh.org points out, "Picking up a saw does not guarantee a straight cut. Holding a chisel does not make you a sculptor." The same logic applies to AI. A tool's usefulness depends on the operator's ability to direct it, interpret feedback, recognize mistakes, and make corrections—requirements that apply just as much to LLMs as to any other instrument.

The False Dichotomy: Tools vs. Automation

Hickey's distinction between tools and automation creates an artificial divide. Automation can be—and often is—a feature of a tool. A compiler performs substantial work without the programmer directing each operation. A spreadsheet recalculates results without manual calculation. Yet we still consider these tools because we wield them by specifying inputs, examining results, and deciding what to do next. Why should generating text automatically disqualify an LLM when translating code or calculating numbers does not?

The rebuttal further dismantles Hickey's claim that LLM use lacks intention, feedback, and expertise. A programmer can ask an LLM to explain unfamiliar code, challenge that explanation, inspect the source, test a change, reject it, and refine the request. That is a feedback loop directed by a human pursuing a purpose. Doing it well requires knowing what to ask, what evidence matters, and when the answer is wrong—expertise that makes the instrument more useful.

The Cognitive Offloading Crisis: Real Risks Backed by Data

While the philosophical debate rages, empirical evidence is mounting that misuse of AI has real consequences. A 2025 study cited by Unite.AI found that frequent AI tool use is linked to weaker critical thinking performance, driven largely by a pattern researchers call "cognitive offloading"—delegating thinking to the tool instead of doing it yourself. The effect was most pronounced among younger participants, raising alarms for K-12 education.

The stakes extend far beyond the classroom. The OECD's most recent Survey of Adult Skills found that adult literacy and numeracy have declined or stagnated across most member countries over the past decade. Adults with stronger numeracy skills are more likely to be employed and earn higher wages. Students who routinely hand their thinking to AI risk accelerating that trend, with consequences for their future earnings and the broader economy.

Two Skill Sets, Taught Together

Responsible AI use requires two distinct skill sets, and schools that focus on only one are setting students up for a false sense of competence, argues Unite.AI. The first is technical: understanding how AI tools generate answers, learning to verify outputs instead of accepting them at face value, recognizing bias and hallucination, evaluating which tools are appropriate for a particular task, and understanding data privacy implications.

The second is human: critical thinking, empathy, adaptability, decision-making, and judgment. These are the skills that help students decide whether to use AI at all and maintain real relationships. The risk is tangible—half of high school students surveyed by the Center for Democracy & Technology said using AI in class makes them feel less connected to their teacher.

Enterprise AI: The Wrangling Problem

In the enterprise, the challenge is less about adoption and more about governance. Hanzel Gregorius of Numa, in an Authority Magazine interview, described the progression: "Once everyone's comfortable, everyone's using it differently, and you get dozens of private approaches inside one company." The challenge becomes wrangling—which tasks should be automated, which shouldn't, where something needs to be checked twice versus three times. That comes down to judgment more than policy.

Gregorius's key insight: "Exploratory work can be roughly right fast. Anything that ends up in front of a customer can't be. Knowing which one you're in is the entire skill." This is the operational definition of AI mastery—knowing when to trust the output and when to double-check, when to delegate and when to do it yourself.

The Cost of Asking for Help Too Late

Gregorius also shared a cautionary tale that illustrates the human side of AI and work. He needed to deliver numbers quickly, avoided stakeholders to move faster, pulled everything himself, and shipped it to clients. The numbers were wrong. "The real cost wasn't the timeline," he said. "I'd gone around people, and putting a number someone else owns in front of a customer without them breaks trust in a way that's much harder to repair than a spreadsheet."

This story highlights that being helpful isn't the same as being the right person. "Having the humility to say 'I am not the best person to do this' is the actual skill." AI doesn't change that fundamental truth—it amplifies the consequences of ignoring it.

The Skills Visibility Gap and the AI Proficiency Divide

HR leaders are waking up to a new set of challenges. According to HR Dive, few CHROs feel very confident they can manage AI-driven change. A University of Phoenix survey revealed a "skills visibility gap"—HR teams have an incomplete view of their employees' capabilities. IBM research found that AI could erode skills like critical thinking and judgment, the very competencies that make humans valuable.

There's also a divergence between employee groups based on AI proficiency. PwC reports that employees who lack specialized skills and do not excel at AI use are at risk of falling behind their peers. Meanwhile, employees may use AI without fully understanding its purpose, leading to a lack of real productivity improvement—a costly failure given the investments companies are making.

The Financial Imperative: AI Skills Pay

Nowhere is the value of AI skill more evident than in finance. A Fortune report notes that as companies invest in AI, automation, and data analytics, they need talent that can translate those investments into results. Finance and accounting professionals who can apply AI tools effectively while maintaining accuracy and compliance will be especially valuable—and better compensated.

The CFO role is expanding to encompass enterprise technology and AI strategy. But as the report emphasizes, technical proficiency with AI is only part of the equation. The ability to apply judgment, maintain compliance, and build teams that can support AI priorities is what truly differentiates high performers.

Conclusion: The Skill of Wielding AI

The debate over whether AI is a tool or automation is a distraction. The useful distinction, as the yyhh.org rebuttal concludes, is between "wielding an instrument with skill and delegating work without adequate understanding or oversight." That distinction applies to AI, and it gives us something more practical than declaring that an entire class of software cannot be a tool.

The evidence is clear: AI is a tool that amplifies both good and bad judgment. It can accelerate learning or accelerate cognitive decline, boost productivity or create false confidence, enhance careers or expose them to risk. The difference lies in the skill of the person wielding it—and that skill is built on technical understanding, human judgment, and the humility to know when not to use AI at all.