Sébastien StormacqThere is a full RAG engine in Kiro-CLI. No vector database installation, no chunking, no dealing with Bedrock models.
It helped me to write the recent S3 anniversary blog post.
Here is how 👇
https://stormacq.com/2026/03/14/how-kiros-knowledge-base-helped-me-write-the-s3-anniversary-blog-post/
How Kiro's Knowledge Base Helped Me Write the S3 Anniversary Blog Post
Today, twenty years ago, AWS quietly launched Amazon S3. What started as a simple object store with two primitives, PUT and GET, became the foundation upon which large enterprises built their data lakes, and with the introduction of S3 Vectors, is now powering a growing number of AI workloads. I had the honour of writing the S3 anniversary blog post for AWS. For that post, I wanted to go beyond the usual narrative and share concrete technical details about the infrastructure behind S3 when it first launched. I interviewed Senior Principal Engineers who were around twenty years ago and collected stories about the early days. But stories alone weren’t enough. I needed data. What was the initial storage capacity? How many objects could the system hold? How much bandwidth was available? To publish numbers like these in an official AWS blog post, anecdotes don’t cut it. I needed evidence.
Qiita - 人気の記事