Can AI Help You Find the Work You Are Actually Suited For?
*This is independent WhatAI coverage of a Chris Koerner episode. It is not career, psychological or financial advice.*
Chris Koerner argues that "follow your passion" is weak advice. His alternative is to examine the evidence already hidden in your behaviour (what you watch, buy, envy, research and return to) and use AI to identify the patterns.
**The idea**
In episode 316 of The Koerner Office, Chris Koerner explains why he believes "follow your passion" is poor advice. The official episode description says he uses eight practical exercises involving YouTube history, spending habits, childhood interests, energy, envy and AI to uncover repeated patterns.
His central argument is that passion often follows progress. Instead of waiting to feel certain about a calling, a person should examine existing evidence, choose a direction and take enough action to discover whether interest deepens with competence.
AI creates a new version of this exercise. A person can give a model a structured record of projects, interests, skills, frustrations and behaviour, then ask it to identify recurring themes.
**Your digital history may reveal more than a personality test**
People often describe themselves using a fixed identity: "I am creative," "I am not technical," or "I have never known what I want." Their behaviour may tell a more detailed story.
A YouTube Watch Later list can reveal recurring subjects. Purchase history can show what someone values enough to fund. Browser bookmarks can expose problems they repeatedly try to solve. Old projects can reveal the kind of work they begin even when nobody assigns it.
AI can process a large set of messy observations and group them into themes. It might notice that apparently unrelated interests share a common mechanism: teaching, designing systems, investigating claims, building communities or turning complicated information into practical tools.
That can be useful, but the model is performing pattern recognition on the evidence it receives. It is not discovering a predetermined destiny.
**A prompt is only as good as the evidence behind it**
Asking [ChatGPT](/tool/chatgpt) or [Claude](/tool/claude), "What career should I choose?" gives the model very little to work with. A stronger process would provide:
- A list of subjects you have repeatedly watched or read about.
- Projects you started voluntarily.
- Tasks that increase or drain your energy.
- Skills other people ask you to use.
- Problems you spend money trying to solve.
- Work you envy and the specific reason you envy it.
- Childhood activities that persisted without external pressure.
- Constraints involving income, location, family, health or risk.
Then ask the model to separate observation from inference, show contradictory evidence and propose several hypotheses rather than one confident answer.
**Where AI can mislead you**
AI tends to build a coherent narrative from the material it receives. Coherence can feel like truth. If you selectively provide evidence supporting a dream, the model may strengthen that dream rather than challenge it.
Viewing history also contains noise. People watch subjects for entertainment, anxiety, novelty or escapism. Spending can reflect aspiration rather than capability. Envy can point toward a desired status or lifestyle rather than the actual work required to achieve it.
Privacy is another concern. A complete history of viewing, purchases, personal projects and emotional reactions can be highly revealing. Users should consider what they upload, whether sensitive records are necessary and how the chosen service handles data.
**The answer should be an experiment, not a label**
The most useful output from an AI career audit is not "You should become a founder" or "You are meant to work in design." It is a short list of testable directions.
For each direction, create a 30-day experiment:
1. Choose one concrete output.
2. Set a small weekly time commitment.
3. Share the work with real people.
4. Measure energy, progress and external response.
5. Record what you enjoyed and what you avoided.
6. Decide whether to deepen, modify or stop the experiment.
Action provides information that a model cannot generate. You learn whether you enjoy the repetitive parts, whether your curiosity survives difficulty and whether the market values the result.
**A practical AI prompt structure**
Instead of asking for a single career recommendation, try a request with this structure:
> Analyse the evidence below and identify five recurring themes. Separate direct observations from your interpretations. Show evidence that contradicts each interpretation. Then propose three low-cost 30-day experiments that would help me test possible directions. Do not assume that a frequently watched subject is automatically suitable as a career.
A follow-up model can then critique the first answer. The goal is not agreement between models. It is to expose assumptions before acting.
**The question for the WhatAI community**
**Could an AI model identify a useful career or business direction from your digital history better than a traditional personality test?**
What evidence would you trust most:
- Watch history
- Past projects
- Spending patterns
- Skills recognised by other people
- Energy and motivation
- Results from actual experiments
Would you be comfortable giving an AI enough personal information to perform this analysis? And has progress ever created passion for something you were not initially excited about?
**Source:** [The Koerner Office, episode 316, published July 10, 2026](https://toolkit.tkopod.com/podcast/episode/1c46)