[Release] weaviate/weaviate v1.38.5

Weaviate's latest release is all about fixing problems rather than introducing shiny new features. Performance improvements for LSM stores and the resolution of batch vectorization deadlocks are solid steps forward. But here's the catch: without concrete benchmarks or metrics, it's tough to gauge how much these changes actually enhance performance. Technical debt isn't just about features; it's the unseen issues that can cripple your pipeline when the clock strikes 2 AM.

What they're not saying is that while these updates may improve stability, the lack of quantitative data leaves a gap in understanding how it stacks up against competitors like Pinecone or Milvus. If you're already using Weaviate, these fixes may be welcome, but if you're evaluating vector databases, the absence of performance benchmarks makes it hard to justify a switch.

The real question is: who should care about this release? Teams already invested in Weaviate who have faced issues with batch processing or LSM operations will likely benefit from this update. However, if you're considering a new vector store, weigh the potential gains against the established alternatives. Without independent performance verification, this release may not tip the scales in Weaviate's favor.

Given these considerations, I'd recommend testing this release if you’re already in the Weaviate ecosystem and have experienced the described issues. Otherwise, the lack of substantial improvements relative to competitors suggests you might be better off looking elsewhere for your vector database needs.