#29 Your AI twin just closed the deal
Hi again,
The line between enhancement and replacement is thinner than we think.
So, let’s jump forward to a fictional scene in 2035 for a moment.
Mateo Alvarez had just unboxed the newest Unitree humanoid model—built initially for robot boxing in the hugely popular “Iron Fist King: Awakening” events—and repurposed it as the ultimate corporate stand-in.
Ironically, the world’s favorite fighting machines were now attending shareholder meetings, boardrooms, and investor pitches.
Mateo’s real expertise? Crafting hyper-optimized human identities using two controversial tools:
- Cluely Ultra: The real-time AI assistant that once scandalously coached job applicants through live interviews.
- Mechanize: The ambitious AI decision-automation system from renowned researcher Tamay Besiroglu, built to replace human decision-making across industries fully.
Enter Adira: an 18-year-old entrepreneurial phenom launching a tech startup.
With investment rounds approaching, Adira requested the full package—Cluely for pitch scripting, Mechanize for behavioral decision-making, and the Unitree bot to deliver it all flawlessly in person.
At first, it worked. But within weeks, Mateo noticed that not Adira, but the proxy, was making news appearances, attending panels, and pitching deals. Adira had quietly outsourced her entire public identity.
Mateo felt a pang of unease. Adira wasn’t alone; many of his clients had subtly crossed from enhancement to outright replacement—human experience quietly supplanted by performance optimization.
Sound too far-fetched? We’re likely already on this path, so back to 2025.
The AI Learning Guy newsletter 🤖 🧠💡
AI learning hacks and mega prompts delivered to your inbox.
Why this matters right now
Mateo’s clients might be fictional, but the tools aren’t. These systems already exist—in early form—and they’re evolving fast:
Cluely Positioned as “your second brain for high-stakes conversations,” Cluely provides real-time language support during job interviews, sales calls, and academic exams.
The company’s mission: “We don’t cheat. We level the playing field.” Founded by Chungin “Roy” Lee, Cluely raised $5.3M shortly after Lee was suspended from Columbia for using his own software.
And, the rebrand from ‘cheating aid’ to performance enhancer didn’t change much under the hood.
Mechanize AI Mechanize calls itself “the universal engine of human replacement.” The goal? Automate (all) cognitive labor at scale—decision-making, negotiation, strategy.
The upside: fewer human errors. The tradeoff: we rethink what “work” even means—and who still gets to do it.
Unitree Robotics Known for affordable, high-mobility robots, Unitree develops quadruped and humanoid models used in research, logistics, and entertainment.
Their recent event, Iron Fist King: Awakening, features a live-streamed humanoid robot boxing match, blurring the line between tech demo and spectacle, and raising new questions about robotics and performance.
Woops! Seems like they are no longer just tools. They’re presence.
What once sounded futuristic is now in product pipelines, pitch decks, and—sometimes—on stage in your place.
So, how to approach these tech developments with a clear mind? Let’s examine these four ideas.
1. The ethical red flags checklist
One option is to use this checklist when considering any AI tool that influences how you present yourself, or who presents themselves as you.
- Privacy: What’s being captured? Who’s using it? Cluely needs conversation data. That’s not neutral.
- Fairness: Who gets access—and who’s left behind? If it’s not evenly distributed, it’s not just a tool. It’s leverage.
- Authenticity: Are you building capacity—or faking fluency? Adira didn’t lie. But she stopped showing up as herself. That’s different.
- Harm potential: What happens to your sense of worth when your unaugmented self feels like a downgrade?
- Autonomy: Are you still calling the shots—or just riding the script? Mechanize doesn’t ask what you want. It predicts what will work.
Your answers don’t necessarily mean they have to be dealbreakers. I’d suggest viewing them as pressure points for now. Touch them—feel what stings.
2. The AI impact evaluation grid
Here’s another no-theory, zero-fluff framework to make sense of your next AI experiment:
Personal alignment
- Does this align with what you believe in?
- Would you be okay if others knew you relied on it?
- Does it extend who you are, or overwrite that?
Societal impact
- What changes if everyone uses it?
- Does it reinforce trust, or bend the rules silently?
- Does it amplify good, or just redistribute advantage?
Practical utility
- Does it actually help, or just look smart?
- Is this sustainable, or a shortcut that backfires later?
- Does it add skills—or just polish?
No score. No winner. Just a mirror.
Mateo didn’t ask these questions. Maybe you do.
3. The integrity scale
A fast way to name your line—so you don’t cross it without realizing.
Level 1: Assist – You lead. AI helps.
Level 2: Enhance – You edit. AI improves.
Level 3: Co-create – You collaborate. AI drafts.
Level 4: Represent – AI acts. You supervise.
Level 5: Replace – AI performs. You disappear.
Adira? Full-on Level 5. Mateo? Hovering between 3 and 4. You? That’s the real question.
I’m sure that most of us probably don’t need complex rules. But a gut-check might be useful.
4. The mindset model: Control → Influence → Acknowledge
Shortcuts don’t mean you skip responsibility. They shift it.
Here’s how to frame your role:
Control: What you use. Where you draw the line.
Influence: The conversations you shape. The norms you help set.
Acknowledge: What others do. What tech does. What markets push.
This framework is based on the fact that you (likely) can’t fight the tide. But you can steer your own boat.
See the newsletter edition before last for more on this topic.
No judgment. Just perspective.
Well, I’m not sending out today’s newsletter to declare what’s right or wrong.
I’m here to refine my perspective and explore different viewpoints with you.
Some tools feel helpful. Others feel like they hollow something out. That’s for you to notice, not me to call out.
So, quick gut check:
What part of your work, learning, or communication do you never want to hand off to AI? Reply here.
If something just popped into your mind, hold onto it. If not? That silence might be worth listening to.
Have a great weekend,
Mark
The AI Learning Guy
👋⚡😎
The AI Learning Guy newsletter 🤖 🧠💡
AI learning hacks and mega prompts delivered to your inbox.
Quick reads and books
- Best AI Books 2025. View on Amazon*.
- Clueley – Business Insider | Verge
- Mechanize AI – TechCrunch | Fortune
- Global Robotic Trends 2025
- Promising AI Startups for 2025 – Analytics Insights
Note: No single website has all the answers. This list serves as a starting point for those who want to explore or satisfy their curiosity about AI. Links: Links with * are affiliate links. See disclosure below.
Affiliate disclosure: To cover the cost of my email software and the time I spend writing these emails, I sometimes link to products. Please assume these links are affiliate links. If you choose to buy through my links, a big THANK YOU – it will make it possible for me to keep doing this.