Surviving AI – Job Automation & Workforce Future Insights
Let's be honest: Your company is already planning how AI will replace you.
They're not evil. They're practical. AI is faster, cheaper, and doesn't need health insurance. The only question is whether you'll see it coming and adapt—or be blindsided like millions before you.
I'm Carlo Thompson, Distinguished Engineer. I've spent two decades building the networks that now power AI. I understand this technology from the inside, and I'm here to translate it into survival strategies you can actually use.
Surviving AI delivers:
✓ Early warning signs your job is vulnerable
✓ Skills that AI can't replicate (yet)
✓ Career pivots that protect your income
✓ Real case studies from the automation frontlines
✓ The truth about "AI will create more jobs than it destroys."
Episodes are 30-45 minutes—no fluff, no filler—just the insights you need to stay employed in an AI-powered economy.
For: Professionals 30-50 in customer service, middle management, marketing, HR, finance, operations—basically anyone who isn't a software engineer.
This isn't fear-mongering. It's a wake-up call. Because hope isn't a strategy, but preparation is.
New episodes on Wednesday. Because by Friday, your competition will have already listened.
---
Created with AI tools (Claude, Notebook LLM, ElevenLabs, Descript) to prove humans and AI work better together than either does alone.
Surviving AI – Job Automation & Workforce Future Insights
Ai Dipatch: Ai Training Make Superhuman Unreliable
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
AI is passing the bar exam, acing medical licensing tests, and crushing coding challenges. So why does research show these same systems fail more than 90% of the time on tasks lasting over four hours?
The answer lies in how AI gets trained—and the limitations that process bakes in from the start.
In this episode of Surviving AI, we go deep on the training problem: the gap between benchmark performance and real-world reliability that creates both risks and opportunities for your career.
What you'll learn:
- How Reinforcement Learning from Human Feedback (RLHF) optimizes AI to sound right rather than be right—and why this creates "sycophantic" systems that tell you what you want to hear
- The "4-Hour Rule": why AI succeeds on quick tasks but struggles with complex, sustained work—and what that means for job vulnerability
- Four predictable failure modes you can learn to spot: temporal blindness, distribution shift, benchmark theater, and inherited bias
- Why 98% of companies feel urgency to deploy AI while only 13% are actually ready—and what happens in that gap
- The three emerging roles that become MORE valuable as AI capabilities grow: the Validator, the Translator, and the Accountability Layer
- Specific questions to ask when AI enters your workplace for hiring, strategy, or workforce decisions
Key research discussed:
- "Open Problems and Fundamental Limitations of RLHF" (Casper et al.)
- METR's research on AI task completion by duration
- Cisco's 2024 AI Readiness Index
- Stanford's Foundation Model Transparency Index
- Deloitte's findings on executive decisions based on AI hallucinations
The bottom line: The gap between what AI benchmarks measure and what work actually requires is your competitive advantage. This episode shows you exactly where to find it.
YouTube Episodes