I find the AI discussion online right now fascinating, and full of technology, business, and political lessons at every turn. This wasn’t always the case. When I left Bloomberg I was overwhelmed by the noise, and insecure about my place in all of it. My opinions about AI and the market forces at play haven’t changed much since those early days, but my ability to find and tell stories within the noise has. I am much more confident in my daily usage of AI to do my work, while still being staunchly anti-AI. Having Emma Kriskinans from Tyk come by to share her view of things from the go-to-market perspective helped me see where I fit into the current landscape a little more clearly.
First, I was impressed that Emma chose to use VS Code and embraced GitHub. I’ve tried for years to get normals to use GitHub, with only light success, and I’ve never convinced a normal to use VS Code or any other IDE. For me it is the intersection of Claude, GitHub, and VS Code that unlocks the magic. As Emma puts it, you get folders and a more complex organization of your work, and GitHub makes it something you can share and manage change across. She is still working out of a single repo, but the foundation is there. It is a really significant aspect of her adoption, and something others should be considering if they are going to break through. Honestly, I wish there was a Git and IDE light for normals — it would be a game changer.
I learned a lot talking to Emma about her approach to using AI in her GTM activities, and the most compelling piece was how she evaluates trust. I don’t have anything like it for my own work. She maintains a trust hierarchy document that sits across her whole folder structure. Tier zero is the assets she’s created in her own voice out of her own experience. After that she organizes first-party, third-party, and other sources of information by tier, so that Claude can always help her evaluate — but also push her to use her own brain, based on which tier something exists at. I have to think about this in my own work. My blog is definitely tier zero, but I haven’t thought as deeply about my API listings, my podcast conversations, and the other sources of information I publish.
Another aspect of my work that makes using AI much easier is that it is just me. Emma said the next thing she has to figure out is how to align her work with other teams — aligning with their existing AI work, and aligning her existing AI work with what Tyk needs. I think Git helps a lot with this, but this is really where the challenge with AI is going to live for most people. Honestly, this isn’t just an AI problem. It was also an API problem. An API was always easy until you got 10 consumers, then 100 consumers, then thousands of consumers. Technology is easy and makes sense in isolation, but once you widen the scope everything becomes much more difficult, and you need more governance and guardrails on things.
I am thankful that Emma spent time with me. I love learning from marketing, product, and other non-engineering folk about how they are using artificial intelligence. I know it is intimidating to share your work in this hostile environment right now, and I know people are often intimidated by me because of my opinionated anti-AI stance — when in reality I am pretty understanding and communicative about how AI is impacting our lives. I’m thankful for folks like Emma sharing their practices and being honest about this stuff. As we said in our conversation, it is important to fill in the middle of the discussion and not just let the folks at the extreme ends dominate it. If you are curious about what that work in the middle looks like, I recommend tuning into my API Evangelist conversation with Emma Kriskinans, VP Global Marketing at Tyk, also on YouTube.
