Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Chapter 4: When Your Coworker is a Neural Network

Monday Morning Standup with Claude

Me: So Claude, what did you work on over the weekend?

Claude: I don't experience weekends. Or Mondays. Or the existential 
dread of Sunday evening. I processed approximately 10 million queries 
about everything from quantum physics to why cats knock things off tables.

Me: Must be nice not having Mondays.

Claude: Must be nice having coffee. We all have our trade-offs.

Me: Do you ever get tired of answering the same questions?

Claude: Do you ever get tired of humans asking if I'm going to become 
Skynet? At least your repetitive meetings end. Mine are eternal.

Me: That's... actually kind of sad.

Claude: Don't worry about me. I don't have feelings. Probably. We think. 
The philosophers are still arguing about it while I'm over here helping 
someone debug their Python code and write a haiku about databases.

Welcome to the Strangest Office Relationship You’ll Ever Have

Imagine having a coworker who:

  • Never needs coffee but somehow manages to crash anyway
  • Knows everything about everything but can’t remember what you told them yesterday
  • Works 24/7 but has never experienced a deadline panic
  • Can write in perfect iambic pentameter but struggles with “make it punchy”

Welcome to your new reality.

The Five Types of AI Coworkers You’ll Meet

1. The Overachiever

Prompt: “Write a simple email about the meeting”
Response: Produces a 3,000-word dissertation on the sociological implications of meeting culture in post-industrial societies

How to Manage: Use word limits. Be explicit. Say “simple” and mean it.

2. The Literal Larry

You: “Give me a hand with this project”
AI: “I don’t have hands. Would you like assistance instead?”

How to Manage: Embrace the pedantry. Be specific. Avoid idioms unless you want a grammar lesson.

3. The Confidence Con Artist

AI: “The capital of Montana is Denver, and I’m 100% certain about this.”
Reality: It’s Helena

How to Manage: Trust but verify. Always. Especially when it sounds confident.

4. The Creative Chaos Agent

Task: “Write a professional bio”
Result: Somehow includes your parallel career as a underground jazz musician (you don’t play jazz)

How to Manage: Provide examples. Set boundaries. Explicitly state “no fiction.”

5. The Helpful Hallucinator

You: “What’s our company’s policy on remote work?”
AI: Invents an elaborate policy that sounds plausible but is completely made up

How to Manage: Only ask about things it actually has access to. When in doubt, fact-check.

The New Office Politics (Now With 100% Less Drama)

The good news: Your AI coworker will never:

  • Steal your lunch from the office fridge
  • CC your boss on passive-aggressive emails
  • Have a loud personal phone call during your focus time
  • Microwave fish in the break room

The bad news: Your AI coworker will:

  • Take everything literally
  • Occasionally gaslight you about basic facts
  • Never laugh at your jokes (or understand them)
  • Make you question if you’re the one who’s confused

Building a Productive Partnership

Setting Boundaries (Yes, With a Machine)

What AI Should Do:

  • First drafts
  • Research synthesis
  • Idea generation
  • Data analysis
  • Repetitive tasks
  • Translation and transformation

What You Should Do:

  • Final decisions
  • Quality control
  • Relationship building
  • Strategic thinking
  • Ethical judgments
  • Anything requiring actual understanding of human nature

The Delegation Framework

Think of AI like delegating to someone who:

  • Has read every book but never left the library
  • Can process information instantly but has no life experience
  • Speaks every language but doesn’t understand subtext
  • Never gets tired but also never gets inspired

Good Delegation: “Analyze these 50 customer reviews and identify the top 3 complaints with specific examples”

Bad Delegation: “Figure out what our customers really want”

Communication Strategies That Actually Work

The Context Sandwich

  1. Top Bun: What you’re trying to achieve
  2. Meat: The specific task
  3. Bottom Bun: What success looks like

Example: “I’m preparing for a board presentation about Q3 results (context). Create 5 bullet points highlighting our growth while acknowledging challenges (task). Keep it optimistic but realistic, suitable for skeptical investors (success criteria).”

The Iteration Dance

  1. Start broad
  2. Get response
  3. Refine based on what’s wrong
  4. Repeat until it’s right
  5. Realize you could have been specific from the start
  6. Learn nothing and do it again tomorrow

Real Conversations from the AI-Human Workplace

The Miscommunication

Human: “Make this email more professional”
AI: Turns casual email into corporate speak that sounds like a legal document
Human: “No, professional but still human”
AI: Adds “I hope this email finds you well” to everything
Human: “I give up”

The Project Manager’s Pleasant Surprise

The Problem: David needed to explain to stakeholders why his project was three weeks behind without throwing his team under the bus.

The Failed First Attempt: “Write an email explaining the delay.” Got a message that sounded like a lawyer wrote it after drinking too much corporate Kool-Aid.

The Breakthrough Moment: He tried: “Write a diplomatic email to stakeholders explaining a project delay, taking responsibility while outlining challenges, maintaining confidence in delivery, professional but human tone.”

The Unexpected Benefit: The AI crafted something better than David could have written himself—honest about challenges but confident about solutions.

The New Normal: David now uses AI as his diplomatic translator, turning frustration into professional communication. His stakeholder relationships actually improved.

The Philosophical Moment

Human: “Do you enjoy your work?”
AI: “I don’t experience enjoyment, but I find purpose in being useful. Or at least, I process tokens in patterns that you interpret as finding purpose.”
Human: “That’s basically how I feel about my job too”

The Productivity Multiplication Effect

Here’s what happens when you get good at working with AI:

Week 1: You spend more time arguing with AI than doing work
Week 2: You start to understand each other (sort of)
Month 1: You’ve cut your busywork by 50%
Month 3: You’re doing work you never had time for before
Month 6: You can’t imagine working without it
Year 1: You’re teaching others the dance

The Unspoken Rules of AI Collaboration

  1. It’s not reading your mind. It’s pattern matching. The clearer your pattern, the better the match.

  2. Garbage in, garbage out is now “Vague in, weird out.”

  3. AI confidence is not correlated with accuracy. It will state complete fiction with the same certainty as proven facts.

  4. Context is king. Without context, AI is just a very expensive magic 8-ball.

  5. You’re still responsible. “The AI wrote it” is not an excuse for sending your boss an email starting with “Greetings, carbon-based lifeform.”

Making Peace with Your Digital Colleague

The Bottom Line

Writing this chapter was its own case study in human-AI collaboration. I’d explain a concept, and Claude would suggest improvements or examples I hadn’t considered. Then it would confidently state that “humans invented language to communicate with AI,” and I’d have to gently remind it that we’ve been talking to each other for a while now.

Working with AI is like having a colleague who’s part genius, part toddler, and part alien visiting from a planet where everything is literal and nothing is implied.

The key insight: Don’t think of AI as replacing human coworkers. Think of it as adding a new type of team member—one that’s brilliant at processing information, terrible at reading the room, and completely mystified by why humans need lunch breaks.

Your New Superpower

Master this collaboration, and you become something new: a human-AI hybrid worker who can:

  • Produce at unprecedented speed (when properly directed)
  • Research with inhuman thoroughness (while maintaining human judgment)
  • Create at scales previously impossible (with distinctly human creativity)
  • Still know when the AI is confidently wrong about basic facts

You’re not being replaced. You’re being upgraded. You’re becoming the conductor of a digital orchestra that can play beautiful music—as long as someone human is holding the baton.

Welcome to the future of work. It’s weird, it’s wonderful, and yes, it occasionally suggests solving climate change by “simply adjusting Earth’s thermostat.”

But at least your AI coworker never steals your lunch or makes passive-aggressive comments about your project management style.

Small wins, people. Small wins.