Good episode covering Delta Lake, Iceberg, and Hudi. The conversation on table format choice was the most useful part — Iceberg’s hidden partitioning is a real quality-of-life improvement over Hive-style partitions you have to manage manually.
Key takeaway: lakehouses close 80% of the performance gap with pure warehouses, while keeping data in open formats you own. The 20% gap only matters for very high-concurrency BI workloads.