Oh, the AI. It’s EVERYWHERE. (But I’m going rogue here and not using AI to generate this content.)
I have so much to say on AI. So much.
I’m a Builder.
And was an early ChatGPT adopter, back in December of 2022.
I’ve used some form of generative AI near daily since then. In the last 18 months, I’ve invested hundreds of hours building with v0, Lovable, Claude Code, n8n, Zapier, Supabase, and more. As a technical, but not engineer human, vibe coding has allowed me to solo build things that previously would have taken a dev team, tens to hundreds of thousands of dollars, and many months. And that was exhilarating for a while.
Eventually, everything started to look the same. And the sense of satisfaction I’d previously feel when digging into thorny tech problems just wasn’t there. There was never an end to the backlog and the adjustments because I could so easily always keep going. Vibe coding started to feel like the infinite scroll. At the end of a session, I’d often feel icky and weird rather than proud and accomplished.
I still build with these tools. Sometimes the act of building itself feels therapeutic, so it’s become another art medium. And I enjoy visualizing the “art of the possible” for myself and for clients. I’ve always been an ideas person, but sometimes I’m so far in the future on my thinking that what I’m describing doesn’t resonate, so vibe coding has enabled me to help concepts better land.
What I’m finding, though, is that we’re forgetting the basics. Everyone keeps chasing the shiny object without actually thinking about how this all might be harmful, why are we moving so fast, and who benefits? Yes, there’s lots of discourse around data centers as there should be, but within the AI builder, enterprise AI, and tech consulting communities, I don’t believe people are paying attention to this enough.
We are also skipping right past establishing team norms and best practices ways of working because #AI. Collectively and globally, we’re pretty bad at humaning right now and it’s not getting better.
I went to a product conference in San Francisco some weeks ago where C-level and very senior people were present both on stage and as attendees from all of the major AI labs, and notable tech companies. Nothing on the agenda centered around AI governance and ethics. This was a huge miss, in my opinion. With the level of AI power and influence concentrated in that venue, it was irresponsible to not highlight those conversations. Instead it was just 100% AI-pilled with no room for any thoughtful criticism.
That event showed me that no one really knows what they’re doing and too many people in positions of great power are maybe not being as intentional as they should be. And it left me wondering, where are my fellow conscientious builders? The people who have actively leaned into learning and building, but who are also actively questioning and worrying.
Enterprise AI
My corporate AI consulting work has highlighted that everything is moving faster and people in the space are tired. I don’t know anyone who is working under less stress because of AI. It usually feels like being in the spin cycle of a washing machine.
Pass the slop from the left hand side.
AI-created agendas, decks that all look the same, analysis that all looks the same, Copilot and Claude tell us what to think when we’re not sure of ourselves, Copilot tells the clients how to respond…
Many of us have become robot reviewers above all else.
And most of us don’t need agents. Almost all of what I see coming through an agent intake request pipeline falls into a few categories: AI-assisted automation, traditional RPA (robotic process automation), a need for better enterprise search across many systems (especially complex legacy ones with sh-t UIs), for Microsoft AI focused shops: a better understanding of current Copilot/Teams/SP functionality, including things like scheduled prompts, SharePoint agents, Notebooks, Copilot in Lists, Planner, etc.
Personally, I’ve fallen into the same trap. Does this need an agent? What about this? Have I finally built an agent? Is this truly agentic? The answer is almost always no.
I’ll share another “build in public” view of the CopilotCopilot I built, but a recent addition includes a “Do I really need an agent?” wizard.
Right now I guess I can best describe myself as a conflicted builder. And I’m looking to connect with other conflicted builders.