---
published: true
layout: post
title: Understanding the Way Things Are
image: https://kinlane-images.s3.amazonaws.com/apievangelist/api-evangelist-images/understanding-the-way-things-are.png
date: 2026-06-29
author: Kin Lane
tags:
  - API Governance
  - MCP
  - OpenAPI
  - Agents
  - Artificial Intelligence
---
A common mistake I see technologists make is that we don't always take the time to understand the way things are before we plow forward with something new. I regularly suffer from this condition, but I don't fall prey to the condition as much as I see others do in the API space. We obsessively dwell on the future in this space, demonize our past legacy, and regularly push forward without any regard to the way things are, because we are the smartest people in the room looking to fix all the problems.

When I threw myself into governance at Bloomberg it was tempting to begin with the ideal set of Spectral rules, but I made sure that I mapped the landscape of APIs to better understand the way things were, rather than just focusing on the way things should be. I am not saying we shouldn't be looking forward. I am saying we shouldn't be looking forward without considering the way things are right now. I think many of us service providers are purely in the business of delivering the future without taking into consideration what people are burdened with right now, and the damage that purely looking toward the future can cause as part of existing business operations.

As I work to make sense of how people I am talking to within enterprise organizations have just gone from 0-60 with MCP deployments, without thinking through what they already have in motion with OpenAPI, governance, documentation, and other things--I can't help but think how damaging forward motion at all costs can be. How much money and time is wasted. And how much more effective teams could be if they just took a moment to properly map the existing landscape and apply it to what is being asked of MCP and Agent Skills rollouts. I understand that leadership has mandated this forward motion at all costs, but I see very few companies translating this opportunity into value add on top of their existing API investments, with most just adding yet another dimension of sprawl to the existing landscape they haven't quite gotten a handle on.

I know it is boring to focus on existing infrastructure. We all want to play with the shiny new toys. We all want to focus on what is new, and look forward to the future. But I think as MCP moves from new and exciting into the old toolbox, there is going to be way more work than teams will be able to handle cleaning up the MCP sprawl, while teams are still struggling to manage the existing API sprawl. With this said, I'm spending a lot of time trying to understand the way things are with MCP so that I can thoughtfully map it to the way things are with the other API patterns already in motion. I am looking to offer some pragmatic solutions as the forward motion of this AI evolution continues to slow, and come back down to reality. To do this, I just focus on the way things are and minimize dreaming about the ideal state. I've learned to accept MCP as the way things are, without forgetting how we got here.