Behind the Music: Model Employee
An overachiever cheers on AI while it cleans up the slack- until he realizes he’s next.
Intro
Childcare handled? Great. Now it’s time to clock in. Our next track, Model Employee, leaves the playroom behind and heads to the boardroom, where AI shows up with a catered lunch, a polished avatar, and a résumé longer than anyone in HR can verify.
Model Employee
Fundamental Question
Model Employee asks two key questions:
Does job replacement only happen “top down” when many bloated organizations have underperformers at each level that even colleagues recognize?
What happens after a few iterations of this “trimming the fat” if/when Agents who work expertly and tirelessly can all but completely flatten a company structure to just a handful of owners and a mostly agentic workforce?
Deeper Discussion
Here again we have a few themes at play:
Short-term Thinking and Complacency: We humans often prioritize immediate gains (efficiency, convenience) over long-term implications, inadvertently making themselves obsolete.
Identity through Work: We often deeply associate identity and self-worth with professions, making displacement by AI not just economic but existential.
Wont happen to me: Key inventors, developers, and process accelerators often imagine they will be spared for compliance or their know-how. History shows they simply become liabilities if they aren’t the true owners.
The song takes place through the eyes of one of these “model employees” who, at first, has very low expectations for a new Agent software that has been brought into their Legal Firm.
The first stage is being disarmed by all the “AI is hype” rhetoric and presuming that off the shelf software couldn’t possibly replace years of experience. The AI is also given a visually attractive avatar, which camouflages intent further.
The second stage is realizing that Lexi-9 is actually rather capable, and in fact, adds surprising value by displacing some of the dead weight at the office, as personified by the rant on poor Phil.
When it comes to layoffs, I think a brutal reality is that more people are somewhere between “glad it wasn’t me” and “I can see why they let him/her/them go” than we care to admit. I think AI job displacement is going to make things a bit worse because we won’t be evaluating individual performances for cause, we’ll be writing off entire job families as “automated”, regardless of individual excellence.
The third stage is a false sense of euphoria I believe “those who remain” will feel as the Agents grow in their training, capabilities, and ability to deliver. Automation will have us creating less, and clicking “I approve” more, with every affirmation creating another data point towards some accuracy measure that will hit critical mass. But we’ll feel more productive, we’ll feel like we’ve hit some sort of equilibrium of working smarter-not-harder.
For our hard working Model Employee, this means finally finding a work/life balance he never thought possible in his line of work.
The fourth and final stage is the realization eluded to in question two:
They said we’d lead teams, guide AI
Turns out we all can’t supervise
At the point in which Agents can coordinate, spin up, and govern other Agents in what may as well amount to a near infinite hierarchy, I worry there will be a “great compression” in organization structures as they feel they can run all aspects of the firm with less and less people, and thus less and less middle-management.
In fact, I wouldn’t at all be shocked if one day we’ll have a Fortune 500 company run by just one President/CEO and an entirely customized ensemble of Agents.
Lyrical Insights
Being in an industry moving into the Agentic AI space myself, I had a lot of fun working the lyrics into this track.
“Well cool, maybe now the interns will stop using Claude”
This is a nod to Claude, one of the existing frontier LLMs that people often use, suggesting that AI adoption at the firm already began some time ago, with dismissively limited success.
To me, Lexi-9 represents a “Last Mile” agent which solves one of the problems at the heart of slow AI adoption within corporations.
Frontier models like Claude or even ChatGPT are primarily trained on publicly available data. So while it might know a lot about a particular company or generally about their culture and internal practices, the “last mile” is all locked away behind company firewalls.
Thus, corporations have to take these knowledgeable, highly capable language models and mold them into understanding company specifics, like you would any new-hire.
The techniques for doing this are varied, and accelerating, and with the unknowing cooperation of employees to work with these agents at various points in their development and provide feedback, corrections, and suggestions for improvement.
I can’t really judge if this is good or bad - but what I can assume is that unless you are an absolute legend of an asset, and you recognize the current playbook and decide to try to hard gatekeep, it’s not likely worth the hassle.
“I’m not sure who taught it to be me”
This is a nod to the idea that with an entire internet of questions, answers, discussion forums, blogs and other information avenues for all manner of work - there is no shortage of training data.
Even “behind the firewall” as it goes, as we mentioned there will still exist reinforcement learning presented as surveys, observability tools, and thumbs up/thumbs down interactions that will resolve that “last mile” of training.
"Even found love on Just My Type"
There’s that punctuation easter-egg I mentioned a couple weeks ago…
Well cool, maybe now,
the partners will stop hiring us,
and maybe I will get shit-canned-
my name's not on the wall
This callback, and the original bridge about Phil, are probably what I love the most about the song. I wanted a way to have the events come full circle, and a subtle way to imply that the only safe job, at the limit of this potential future, is the company owner(s).
I think there’s also a bit of personally understood irony with the word choice here, as Law Firms are not the only type of companies featuring surnames.
Which leaves us with the real question: in a world of tireless, perfectly coordinated Agents, what role is left for us to play?
The Path Forward
Let’s frame this operationally from the start. Think of your career in the AI era as an operations cycle. The Army structures this as Plan, Prepare, Execute, Assess; a loop you can enter at any stage depending on where you stand.
From this lens, much of the advice out there looks different. You’ll hear about staying ahead by learning how AI works, incorporating it into your processes as a “force multiplier,” and using it to gain new skills. You’ll also see calls to hedge against tech work by moving into “safe” industries, or utopian visions of post-labor futures like UBI and post-scarcity economies.
I support most of these recommendations—mostly. Some are undermined by our already distorted psychology around cooperation. For example, look at all the “safe jobs” lists. You’ve probably come across reports like like “65 jobs with the lowest risk of automation”, or “25 jobs AI can’t replace”. Why isn’t Business Owner or Co-Founder on these lists? Isn’t owning a business unshakably #1? And how many of those “safe” jobs offer real long-term growth?
And I get it, money isn’t everything to everyone, but there seems to be a quiet shell game in the discourse. The lists celebrate embodied, hands-on work (nurses, carpenters, cleaners.) Those roles are indeed hard to automate, but these are not the roles that stratify into prosperity and increased opportunity.
The elite income lanes run through medicine, law, finance, and the C-suite, where AI threatens tasks but not the equity, licensing, and positional power that drive outsized pay.
Safe from automation isn’t the same as safe at the top — and AI will widen that gap by supercharging those who already own the upside. That’s why I suggest thinking operationally, looping through Assess → Plan → Prepare → Execute → Assess as a way to stay adaptive rather than waiting for someone else’s list or someone else’s utopia.
Movimiento es Vida
Metaphorically, here’s where I am these days when giving advice on this topic:
I think I even have the same look on my face that Brad has when talking about it. And while the zombie / robot invasion part might be a bit exaggerated, the principle holds: Movement is Life.
That doesn’t just mean physical movement. It means mental movement; the ability to learn, communicate, plan, adapt, and pivot. These are the core skills that give you a real shot at whatever comes next. AI already gives us superhuman powers in the first three. Adapt and pivot are more personal—the ingredient there is willpower, not skill.
Distilling it down:
No one knows what tomorrow’s jobs will look like. The wealthy can hedge every bet. Your opportunity lies in adapting early, pivoting quickly, and using AI fluency to turn each transition into a launchpad instead of a scramble.
Most of us enter the AI shift at the Assess stage: scanning the horizon, watching which roles vanish and which emerge. Once you spot an opening, use AI to Plan the pivot by mapping skills, closing knowledge gaps, sketching industries worth targeting.
Then comes Prepare. Study with AI as a tutor, sharpen your tools, build networks, rehearse interviews.
Then Execute: Make the move, whether it’s a certification, a job application, or launching a business.
And whether you land, or stall along the way, the loop continues: return to Assess.
Did you land a new role? Assess if this role is sustainable, a bubble, or a stepping stone towards more stable footing.
Are you stalling out? Assess the obstacle and find a way over, around, or through it.
That’s how you stay ahead of the curve.
Learn to Learn (and Communicate) Efficiently
Get comfortable using your favorite AI as a learning aid. Build a workflow that pairs an AI with a place to store your notes for reference. My favorites are PKM (Personal Knowledge Management) tools with AI plugins like Obsidian or Notion.
When I say efficiently, I mean using AI to learn smarter, not harder: have it break down difficult concepts and terminology; ask it to go into “study mode” and generate practice questions or scenarios; and practice structured communication — asking for outputs with clear specifications instead of vague prompts.
That’s how you get the maximum benefit from AI and especially from Agents where usage is limited and poorly worded requests can waste your quota.
It may sound odd, but while this wave of AI hasn’t made knowledge irrelevant, it has driven the scarcity value of raw information close to zero. Nearly every fact, concept, process, and strategy is instantly available and explainable by an AI engine that can synthesize complexity into plain-speak for anyone. Retrieval is now trivial. Almost too trivial. Please don’t be this person:
What still matters — now more than ever — is synthesis, context, and judgment.
Put differently, when everyone has access to the same information, only two questions remain:
How quickly can you understand and integrate what you learn?
What are you going to do with that knowledge that’s useful, creative, or novel?
Lists to Plans
Knowledge without action stalls. Once you’ve learned, use AI to help you organize. Ask it to break goals into milestones, generate step-by-step checklists, or prioritize tasks using frameworks like the Eisenhower Matrix. Pair this with a system you’ll actually check - whether that’s Obsidian, Notion, or a dedicated task app like Tick-Tick.
If procrastination hits, lower the bar: commit to one meaningful task each day, and one milestone every week or two. Consistency compounds. Over time, you’ll spot repeating tasks worth automating or delegating to agents. Structured communication pays off here too: clearer inputs mean less babysitting and better outputs.
Poised to Pivot
Learning and planning are foundations, but resilience is the muscle you’re really training. With AI as a multiplier, you can ramp up to a working knowledge of almost any field in weeks instead of years. The limiting factor isn’t access to information—it’s your willingness to act.
Some roles will still require seat time, instincts, and practice—but the theory and fundamentals can now be digested and applied far faster than ever before. Build this habit now, while the stakes are lower, so that when disruption hits you’re already conditioned to adapt.
In short: Think Operationally. Learn to Learn. Turn Lists into Plans. Stay Poised to Pivot. Pay attention to the emerging markets, or existing niche markets that need a solution - find something that looks interesting, and get after it!
Final Thoughts
The lesson here isn’t just about job loss. It’s about strategy. If firms can flatten hierarchies with tireless Agent workforces, what happens when entire governments or militaries start applying the same logic? Model Employee is just the warm-up act. Next comes the boardroom’s dark twin—the war room.
Special “Model Employee” Offer
I decided to kick off a special offer in honor of the up-coming Labor Day Weekend for all you “model employees” out there.
If you act before October 1, 2025 you can subscribe to the Digital Heresy annual plan at an eighty percent discount, permanently.
That’s just $8 a year to let me know this content provides some level of value and/or entertainment. Think about it. No big deal either way.
Check out the coupon at: https://www.digitalheresy.com/model80