For years, AWS customers have asked for Savings Plans that cover managed databases. Database Savings Plans promise up to 35% savings for serverless deployments and strong discounts for provisioned instances. You keep flexibility across engines, instance families, sizes, Regions, and deployment models while locking in predictable spend with a 1 year commitment measured in dollars per hour.

You may know the parable. A departing CEO hands the incoming leader three envelopes to open in times of crisis. The first says, “Blame your predecessor.” The second says, “Reorganize.” The third says, “Prepare three envelopes.”

If you’ve been eyeing AI but want privacy, control, and the joy of tinkering, 2025 is the year to build your own stack at home. Thanks to open-source distilled models, GPU acceleration, and a maturing ecosystem, you can run powerful language and image models locally no cloud required. Here’s a practical guide to the best tools, how they fit together, and where to start.

I wanted the next train, platform, and a simple signal for problems without opening a web page. This is a SwiftBar/xbar plugin written in Python that queries National Rail via Huxley2, switches routes based on time, and color-codes the menu bar. It also run in a cron on my homelab during the days and windows where I commute to and from Work.

I built a Python toolkit called IBM Spend Sleuth to help teams analyze IBM Cloud billing data, especially when retroactive resource tagging is not available. Problem it solves: IBM Cloud does not support retroactive tagging. Cost allocation gets messy when names are inconsistent. Spend Sleuth uses flexible pattern-based filtering across any CSV column to group costs, validate months, and turn planning assumptions into reports.