Speakers from IBM, Meta, and Netflix at the All Things AI conference in Durham on April 4, 2026 argued that agentic AI increases developer preparatory work rather than eliminating it. They described issues including context rot, the need for orchestration and decomposition, and highlighted IBM's new mellea.ai library and 'switch brains' research. The talks imply practitioners must invest in context engineering, modular functions, and guardrails.
Key Points
- 1Demonstrates that agentic AI increases developer workload through context prep, orchestration, and ongoing review.
- 2Highlights that context engineering and decomposition are essential to reduce hallucinations and maintain agent accuracy.
- 3Advises engineers to build modular functions, orchestration layers, and guardrails, accepting a recurring 'prep tax'.
Scoring Rationale
Timely coverage of credible conference talks from IBM, Meta, and Netflix provides high relevance, scope, and actionable guidance; the mention of IBM's mellea.ai adds authority. Score limited slightly by modest novelty and lack of deep technical detail.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
