Systems | Development | Analytics | API | Testing

The Model Context Protocol (MCP) is quickly becoming the standard for how large language models connect to enterprise data. As adoption accelerates, engineering teams face a foundational decision: build a custom MCP server from scratch, or adopt an AI data gateway that ships with MCP support, security, and governance out of the box. Both paths have real tradeoffs. This post breaks them down so you can make the right call for your stack, your team, and your risk profile.

Every enterprise moving AI into production is about to face a familiar problem in an unfamiliar form: the cost explosion, but for LLMs. This is *very *similar to what happened with cloud. In the early days of cloud, teams spun up infrastructure with no visibility into who was consuming what. Finance got the bill. Engineering got the blame. No one had the data to make good decisions. It took years of hard-won FinOps discipline to fix that. LLM spend is on the same trajectory *and moving faster*.

Insurtech is picking up pace fast. And it’s not only because of new tech coming in, but also because people today simply expect things to be easier, quicker, and more transparent. That said, technology is still doing most of the heavy lifting here. We’re already used to apps simplifying our everyday lives, so when insurance starts doing the same, it naturally feels like the right move.‍ So, what is Insurtech? And why is it expected to reach $152.4 billion by 2030?

Audit logs are essential for making AI systems accountable, reliable, and compliant with regulations. They act as a record-keeping system, documenting every critical interaction within an AI system, such as user prompts, model decisions, and policy enforcement. Here's why they are crucial: Audit logs are not just a legal requirement - they are a key part of managing AI systems effectively and minimizing risks.

One of the beauties of the Rails framework is the ability to utilize Ruby on Rails associations in your models. These associations allow you to access collections of records in your code with pleasant syntax, abstracting away the need to write underlying SQL queries. That abstraction holds as long as all your data lives in one place. The moment your tables are spread across separate database clusters, certain association types stop working.

Artificial intelligence is transforming the global software industry. Some analysts refer to this shift as a “SaaS apocalypse,” with traditional software companies losing over a trillion dollars in market value. Historically, software vendors executed client visions by writing code. Now, as clients articulate their needs and AI generates code, the industry faces a critical question: What role remains for software vendors? This requires a fundamental shift.

On March 31, 2026, security researcher Chaofan Shou noticed something odd: the complete source code of Claude Code — Anthropic's flagship AI coding CLI — was sitting in plain sight on the public npm registry. 512,000 lines of TypeScript. 59.8 MB of source maps. Everything. The irony? The code contains an "Undercover Mode" specifically built to prevent internal Anthropic secrets from leaking into public commits. They built a secrecy subsystem, then accidentally published everything.