Bob TV Episode 4: Why generic AI fails at mainframe scale - and what works
2 asistentes de 122 grupos organizando
Organizado por IBM Developers Montevideo Meetup
Detalles
Join us for the fourth episode of Bob TV, the monthly digital event series hosted by Max Jesch, IBM Bob’s outbound product manager. Each month, Max sits down with a technical guest to explore real examples of your trusted AI assistant in action—showing how Bob accelerates enterprise software development across the entire lifecycle.
Featuring Special Guest: Ethan Mendel, Development Lead, IBM Z premium package for Bob
Ethan will discuss why mainframe modernization is hard — and why AI alone is not enough.
- The real scope of the challenge: 10-year modernization timelines, skill shortages, and technical debt that runs so deep some clients haven't touched certain programs in decades
- Why generic AI falls short: LLMs achieve only ~30% accuracy on non-trivial impact analysis questions for large mainframe applications — hallucination gets worse as scale increases
- The Premium Package for Z approach: static code analytics + metadata-enriched model intelligence as the foundation, not a replacement for AI but a way to make AI reliable at enterprise scale
Bob TV is your monthly window into Bob capabilities:
- AI assisted development workflows
- Enterprise modernization patterns
- Agent based software engineering
Host: 

Ethan Mendel, Development Lead, IBM Z premium package for IBM Bob