How to Practice Your Job: A Skills-First Approach in the Age of AI

The modern workplace is at a crossroads. The traditional career ladder, built on job titles and predefined roles, is becoming obsolete. To thrive in the AI-powered future, we need a fundamental shift in how we approach our professional growth.

[00:00.672]

[Speaker introduces the concept of shifting from a jobs format to a skills format.]

“We need to move from a jobs format to a skills format for our roles and our career growth. And no one’s ready to talk about it.”

This is the central challenge: reframing our careers around a portfolio of improvable skills, rather than a sequence of job titles. The key to this transformation lies in learning how to practice our jobs, preferably with the help of Artificial Intelligence.

The Training Gap for Knowledge Workers

[00:18.432]

[Speaker explains the inspiration from a Tyler Cowen blog post.]

Inspiration for this shift comes from a 2019 blog post by economist Tyler Cowen, who observed a curious discrepancy in the professional world. Athletes train, musicians practice scales, and performers rehearse constantly. Yet, knowledge workers—the thinkers, creators, and strategists of our economy—don’t really train. There is no equivalent to shooting free throws for a product manager or an engineer. We simply perform our jobs live, every day.

[00:35.392]

[Speaker questions what it takes to practice knowledge work like a pianist practices scales.]

This raises a critical question: what would it look like to practice knowledge work? How can we deconstruct our professional abilities into trainable components, similar to how a pianist practices scales? In the age of AI, we finally have the tools to address this gap and rethink skill development from the ground up.

Breaking Free from the Job Title Trap

[00:54.672]

[Speaker discusses how skills are traditionally tied to jobs in software and hiring processes.]

For too long, our understanding of skills has been inextricably linked to job descriptions. This mindset is so deeply embedded that it’s “literally baked into our software.” Hiring platforms, compensation tools, and promotion ladders all start with the assumption that a specific set of skills belongs to a specific job title. This framework makes it difficult to imagine skills as independent, transferable assets.

[01:26.552]

[Speaker describes the future of work where skills are acquired and used with AI, measured by outcomes.]

Yet, we are rapidly heading toward a world where skills exist independently of roles. In this new paradigm, professionals will be measured not by their title, but by their outcomes—their ability to leverage a unique skill set, in collaboration with AI, to produce meaningful work. Your value won’t be defined as a “Product Manager” or an “Engineer,” but by your demonstrated ability to drive results with your specific talents.

Deconstructing Skills: From Piano Scales to Knowledge Work

[02:02.932]

[Speaker uses the analogy of fractal skills in piano playing to explain skill decomposition.]

To understand how to practice knowledge work, we can look at how physical skills are developed. A skill like playing the piano is fractal; it’s composed of numerous smaller sub-skills. These include the basic finger movements for playing scales, the precise pressure applied to the keys, and the speed and rhythm of the performance. Each of these components can be isolated, practiced, repeated, and improved with feedback.

[02:35.432]

[Speaker talks about the need for narrow situations with repeated feedback for knowledge workers.]

For knowledge workers, the path to improvement is the same. We must find ways to create narrow, repeatable scenarios where we can receive specific, targeted feedback. The traditional method of simply doing our job is akin to a constant “live performance,” which is an extremely inefficient way to learn. We rarely get the chance to practice a high-stakes decision or a critical presentation in a low-risk environment.

The Role of AI as a Personal Coach

[02:58.212]

[Speaker highlights AI’s potential to provide custom feedback for practice.]

This is where AI changes everything. We now have an unprecedented opportunity to create these practice loops. AI can provide the custom, scalable feedback that was previously impossible to achieve without hiring an army of personal coaches. The failure to train isn’t because knowledge workers are lazy; it’s because our work environments are structurally designed against it.

[03:28.172]

[Speaker outlines three reasons why practicing knowledge work is difficult.]

Three main obstacles have historically stood in our way:

1. Fuzzy Outcomes: Unlike a basketball shot that either goes in or misses, the success of a strategy document is subjective and multi-dimensional (e.g., speed, quality, political alignment).
2. Delayed & Noisy Feedback: The impact of a decision made in Q1 might not be clear until Q3, by which time countless other variables have changed.
3. Low Repetition: We don’t get to write hundreds of consequential strategy memos a week, making it hard to build muscle memory through repetition.

The Five Core Skills to Practice in the AI Era

[05:07.412]

[Speaker lists five key repeatable skills for the AI age.]

To begin practicing, we must first identify which skills are repeatable and trainable. Five core skills consistently emerge as crucial for knowledge work today:

1. Judgment: How you frame decisions, define options, and make choices under uncertainty.
2. Orchestration: The ability to transform ambiguous goals into concrete, actionable workflows for both humans and AI.
3. Coordination: Skillfully guiding groups of people through ambiguity to achieve a common goal without creating more chaos.
4. Taste: Possessing and articulating a meaningful quality bar for your work, whether it’s product design, writing, or strategy.
5. Updating: The discipline of changing your mind and adapting your plans as new evidence and context emerge.

These skills are not abstract adjectives for a resume; they are demonstrable patterns that appear in the artifacts we produce, such as decision docs, project plans, and presentations.

A Practical Framework for Skill Development with AI

[07:37.032]

[Speaker explains how AI can be used as a tool to improve professional skills.]

AI is not a magic brain; it’s a powerful tool that can read text, follow instructions, and apply a rubric with tireless consistency. This makes it the perfect practice partner. The process is straightforward:

1. Define “Good” with a Rubric: Work with your team to create a clear, concrete rubric for a key artifact, like a decision document. Ask, “When you say this is good, what do you mean specifically?” Break it down into measurable components (e.g., “decision is stated in one sentence,” “at least two real options are presented,” “risks and tradeoffs are explicit”).
2. Create Annotated Examples: Gather 3-5 real examples of that artifact and mark them up according to your new rubric, explaining why certain parts are strong or weak.
3. Train Your AI Coach: Feed the rubric and the annotated examples to an LLM (like Claude or ChatGPT).
4. Start Practicing: Now, you can create drills. Take a messy, real-world situation (like a long Slack thread) and practice creating a clean, one-page decision doc.
5. Get Feedback: Run your practice document through your AI coach. It will score your work against the rubric, quote the specific parts it’s reacting to, and suggest edits to improve your score. This is your “film review.”

From Individual Practice to Team Excellence

[13:27.427]

[Speaker discusses how to apply this practice framework at a team level.]

This framework is even more powerful when applied at the team level. As a manager, you can lead your team in creating a shared rubric, which builds alignment on what quality looks like. You can then implement an automated AI review process as a first pass on all documents, providing immediate, consistent feedback. By tracking scores over a quarter, you can measure real improvement in the team’s collective skills.

This skills-first approach also revolutionizes hiring. Instead of asking abstract behavioral questions, you can give candidates a short, realistic take-home exercise based on the same rubric your team uses for development. This allows you to evaluate them on the exact skills they will need on the job.

The goal is to get better. The goal is to become useful.

This system encourages transparency around AI use. The focus shifts from policing tools to evaluating the quality of the final outcome and the thinking behind it. By adopting a practice mindset, we can transform ourselves from passive performers into active athletes of knowledge work, using AI as our coach to continuously level up our most critical skills.