Description
Apparently one of the new job interview questions is what you do when it’s 5:00 p.m. on Friday and you still have 20% of your monthly AI subscription left.
My answer?
I go home.
But according to this new version of corporate productivity culture, unused AI tokens are apparently wasted potential.
Employees are becoming “managers of compute,” and leaving Claude Code or other AI capacity unused means you could have been generating more code, running more audits, analyzing more data, building more internal tools or otherwise finding something — anything — to keep the machine busy.
And that’s exactly the problem.
Using more resources does not automatically mean creating more value.
In this video, we’re talking about AI productivity, Claude Code, AI-token usage, employee monitoring, keystroke tracking, mouse tracking, Microsoft Teams status, lines of code, Goodhart’s law and why corporate productivity metrics so often end up measuring activity instead of actual results.
If one employee uses a million AI tokens to solve a problem and another solves the same problem with a thousand, the first employee is not automatically more productive.
If one programmer writes 5,000 lines of unnecessary AI-generated code and another programmer solves the problem by deleting 2,000 lines, counting output volume can make the worse employee look better.
The same thing happens when companies measure mouse movement.
Employees learn to move the mouse.
Measure online time?
Employees learn to stay online.
Measure AI-token usage?
Employees will learn to burn AI tokens.
Once a metric becomes the target, people start optimizing for the metric instead of the actual goal.
And that’s how you get productivity theater.
AI was supposed to help companies become more efficient.
Instead, corporate management may somehow find a way to turn it into another scoreboard employees have to max out just to prove they’re working.
What matters more to you: how much activity someone produces, or whether they actually accomplished the goal?
🔥 Resume reviews, custom resumes, consultations, courses and career resources:
https://joshuafluke.store
#ArtificialIntelligence #WorkplaceCulture #Productivity
👊 Join the community!
https://discord.gg/Joshuafluke
Socials🤳
https://www.tiktok.com/@joshua_fluke
https://www.instagram.com/joshuafluke/ 📸
https://twitter.com/joshuafluke 🐦
📧 Email me directly!: grindreel@gmail.com
📧 Business inquiries: Joshuafluke@thoughtleaders.io
My Gear ⚙️: https://kit.co/JoshuaFluke
❤️ Support my content
https://www.patreon.com/joshuafluke
Chapters:
00:00 The New AI Job Interview Question
00:21 Your Leftover AI Tokens Are “Wasted”
00:45 Employees Are “Managers of Compute” Now
01:38 Activity Is Not Productivity
02:15 Why 100% Utilization Makes No Sense
02:39 The Lawn Mower Example
02:55 Nobody Works at 100% Capacity
03:19 Why Unused Capacity Can Be Necessary
03:44 Keystrokes Are Not Results
03:59 Mouse Tracking, Teams Status & AI Tokens
04:34 Goodhart’s Law Explains Everything
04:59 How Bad Metrics Ruin Call Centers
05:16 Lines of Code Are a Terrible Productivity Metric
05:40 AI Usage Can Punish Better Employees
06:11 Token Maxing Creates AI Slop
06:44 AI Output Creates Hidden Costs
07:24 Unused Resources Aren’t Automatically Wasted
07:49 Maximum Capacity Is Not Maximum Value
08:14 AI Has Diminishing Returns
08:42 Stop Inventing Work Just to Look Busy
09:15 When Does Management Become Responsible?
09:31 Productivity Metrics Create Productivity Theater
09:55 Corporate AI Has Become Corporate Bureaucracy
