Future Ready Leadership With Jacob Morgan Claude Mythos Leaked, AI Agents Done Wrong, JPMorgan's New Performance Rules, and the Gen Z Reality Check
Mar 27, 2026
An accidental leak of a supercharged AI model raises cybersecurity and enterprise risk questions. Leaders warn that AI agents are being built as coworkers instead of workflow tools. A major bank now makes AI adoption a measurable performance requirement for tens of thousands of engineers. A new survey highlights Gen Z's push for both logout culture and seats in the C-suite.
AI Snips
Chapters
Transcript
Episode notes
Train For Fundamentals Not Specific Models
- Focus upskilling on fundamentals and adaptable skills because models change faster than training programs can keep up.
- Morgan warns organizations often use multiple models and updates arrive frequently, making model-specific training short-lived.
Naming Agents Causes Local Optimization Problems
- Senior HR leaders warn against treating AI agents as coworkers with names, org slots, and titles because it drives optimization at the wrong level.
- IBM found naming agents like Harry and Sherlock led teams to focus on individual use cases instead of enterprise workflow design.
Redesign Workflows Instead Of Slotting In Agents
- Treat agents as tools to redesign workflows from first principles rather than plugging them into existing org charts.
- Morgan compares this to 1990s websites that were just digital brochures instead of reimagined interactive services.