Maybe this helps you put "migration pressure" into perspective. EVE Online, the objectively best game to ever exist, has been dreading a python2->python3 migration for two decades. Profitable the whole time in between. You don't need to upgrade your stack every few weeks.

The complexity of adding an MCP to your product (if you already have some sort of API) is a single prompt. Here is precisely what I would dictate into my prompt field today. Maybe this gives you some insight into how I prompt agentic systems: "Add full MCP capabilities to the product, aiming for feature parity with our existing API implementation. For authentication purposes, use the OAuth implementation that is closest to our existing authentication stack. If you have to create authentication confirmation pages, keep the design as close as possible to our login and registration pages as you can. For documentation, create solid, self-contained documentation blocks for every single tool that mirrors the complexity of API docs. Create MCP management, analytics, and log features analogous to our API implementation and draft MCP landing pages that also fit into our existing marketing and customer acquisition scheme. Do extensive web searches around the quirks and expectations of OpenAI and the Anthropic MCP and plugin stores. As we will likely want to submit this MCP to be a first-class tool in either of these and more store systems, pay special attention to how tools need to be configured, described, and tagged to make sure that this will pass plugin store review without issues. Implement the equivalent of the tests we have for the appropriate API routes for this MCP implementation, making sure that we follow protocol implementation best practices. Create a dedicated landing page for agentic website visitors that has an easy path for them to quickly implement the MCP within their own runtime and products. When in doubt about any step of the implementation, please use the ask user question tool to get my feedback. First create a scope document and have me verify it, then implement the full document right away." (Dictated but not read 🤣 using WhisprFlow) This should get you 95% of the way there. The rest is some manual testing, deploying, and telling your customers you have an MCP now.

In fact, you will hit an effectiveness ceiling within hours if you don't understand how software systems work. And it goes beyond being able to read the code. It starts with configuring the harness, clearly expressing what needs to be built in the prompt, and setting up validation and testing systems to ensure functionality is maintained over time. All of these require domain knowledge in programming (plus the actual domain expertise of the field that the solution exists in).

I've been noticing that in myself too, and I feel that, in some ways, I've learned a better, more precise way of expressing a concept I would otherwise have misnamed. Kind of a good thing, right? "gate", "canonical", "pinning", and a few more.

That’s why enterprise has an expectation for MCP and API affordances for every software tool they use. I’m getting these requests every single time for Podscan. It’s a requirement now. People want to have their thing use your thing. And they always already have their own thing.

If there's one thing I really love about AI coding, it's that it makes it possible to build hilariously meta projects like this one (and every single bidding website out there) in minutes. You can just do things, no matter how bonkers and useless they are :D

Bidding websites are novelty projects on the surface. The underlying mechanism is one of raw attention, both on the website itself and in the meta-conversation around it. If you still believe advertising doesn't work, then you're missing the biggest learning from Outbid.

The main lesson founders should take away from this is that the confluence of having built an audience for many years and daring to launch new ideas in public tills the soil for events like this. Maybe ask Jonathan how many failed products he has built before today.

From GitHub's outage post-mortem. "Tell your agents not to commit too often" will be the new "turn off your lights when you leave the room." Honestly, though: this is a terrifying traffic growth pattern when you're already the biggest player. Explains a lot. Thanks, AI!

That feeling when you run a podcast business and your favorite living philosopher (who you read zealously at uni) sends you a casual email about becoming a customer with their pod. Podscan has a lot of well-known customers. But I really have to hold back the inner fanboy today.

The fact that a major part of mitigation was 403ing Copilot requests says it all. Even “just” smart autocorrect now operates at a scale that breaks the largest players when a tiny thing isn’t perfect. Feels like the internet is getting even more brittle with agents & AI tools.

I loved being a developer before AI. And now, I love it even more. I’m as close to being a wizard as humanly possible. I can evoke a magical familiar to do my bidding at any point. How could I not be amazed by this. Every day, new heights. I put on my robe and wizard hat 🧙