David Linthicum, Deloitte Touche Tohmatsu, and Lior Gavish, Monte Carlo Data | Supercloud 4
summary
Argues that enterprises adopt a balanced hybrid cloud strategy with only about 20% truly all-in cloud, and says generative AI succeeds only when enterprise data is integrated via RAG or fine-tune approaches. Describes Monte Carlo-style observability for data pipelines to catch hallucinations, cites finance LLMs (10-Q generation) as low-risk starter use cases, contrasts cloud training economies with on‑prem/edge inference, and reports real cost savings from optimizing Snowflake/Databricks.
AI-written summary · about
chapters (13)
- 0:00Welcome and introduction
- 1:01Why cloud computing guide
- 2:47Hybrid cloud reality
- 3:51AI as a data accelerator
- 6:24Data transparency in AI
- 7:29Automating last mile processes
- 8:55RPA disruption with GenAI
- 11:21AI model specificity debate
- 14:13On-prem vs cloud for AI
- 17:28Cloud operating model limits
- 21:47Cloud cost optimization
- 24:12AI workload efficiency
- 26:25Strategic AI investment
