AI Adoption Patterns

Sponsored by

Happy Summer Solstice. It’s June 21st, the longest day of the year! And it’s my birthday! Last week, I was low on motivation, so I didn't send out a newsletter. But this morning I started to feel some late-newsletter remorse, so why not work a little between back-to-back parties, gifts, and endless adulation? ;)

Part of my delay was my new venture. Karen Kelly and I have built Launch By Lunch. It’s been a wild ride. We launched our first AI-Accelerator six weeks ago. It sold out, and demand forced us to offer more.

We are on our 6th week this week. This week’s newsletter contains many of the key learnings from these sessions. Having 150 non-technical founders build with AI is an astounding experience.

Karen and I have learned a great deal from our session.

First, we have come to expect that 20% of each cohort will build something that can scale into a revenue-generating business. Might not be investable (now), but their ideas have the potential to advance their careers and impact their lives. Rock on!

Second, each cohort has some stealthy founders. We only learn about them post-program.

Third, we both learned very, very quickly that a community was needed. It was asked for, both in sessions and in offline requests. As Karen says, being a solo founder is both expensive and lonely.

We are launching the Inner Circle this week. It’s a membership community. We bring together fractional experts, investors, and founders to support the efforts of our members.

Developer Stressors

Here is what I am learning from senior developers as the industry comes to terms with this game-changer. I write about these issues, not with ridicule and scorn, but with compassion. AI is overturning a lot of processes and existing stakeholders, and it’s making a lot of people who have invested years into their crafts feel uneasy.

I feel the stress myself!

1. Developers who fail with Vibe Coding

One of the most surprising patterns I’ve seen is that the more experienced the developer, the worse their initial outcomes with AI tools. After hearing the seventh senior dev tell me flat out that “AI doesn’t help” — maybe a marginal gain here or there — I started to question whether deep technical experience might actually hinder early AI adoption. My working theory is that senior developers bring too much a priori knowledge to the prompt. They overthink, over-prompt, and over-structure. They know too much — and it holds them back.

In a few cases, I’ve coached experienced devs to “think like a newbie.” That shift helps. Once they stop trying to control everything, they begin to see where tools like Lovable can speed them up. I’ve noticed that this journey comes in phases. Each phase has its language. The moment I hear a dev say, “Wow, there are things I should never do with code again,” I know they’re about to hit an inflection point in productivity.

Not all of them make it that far. Some hit a wall and push back hard. I’ve received troll comments on posts and snarky direct messages. It’s not personal. It’s fear. Fear of changing how we think, how we build, and how we define value as developers.

Caveat: It's not nice to tell these developers that non-tech people have no issues with this. Why can’t you do it? You should be amazing. In a few cases, I have to push back hard and point out that all these vibe coders are lying.

2. Legacy Skills vs Essential Skills

After dozens of conversations with developers — some convinced AI will fade, others quietly anxious — I started to notice something deeper than skepticism or fear. Most of us, including me, don’t have a clear sense of what value we actually bring. For years, our worth was wrapped up in the difficulty of code. Building things was hard. It took time, precision, and persistence. That struggle was the job.

Now the ground is shifting. The uncomfortable question hanging in the air is: What’s essential? And if the core of what I do gets reduced, will the part that remains still justify my salary? That’s not an easy question to face. No one likes watching 80% of their resume get stripped out by a new paradigm.

But it’s real. I have a gut feeling that a massive recalibration is coming. Developers will need to shift from builders to solution designers — people who understand business value, not just syntax. That shift is going to be painful. And it’s going to eliminate a lot of roles that were built around complexity, not clarity.

3. Overestimated Complexity

Another subtle but powerful insight: developers with deep knowledge often inflate the complexity of a task, not on purpose, but because we instinctively recall everything that goes into doing it "the right way." This week I needed to add a set of cloud functions to a Vibe app. My default move? Spin up an endpoint on a cloud provider, wire it in manually, and then integrate it with the app logic. That’s what experience taught me to do.

But I paused. I asked myself if I was overthinking it. Then I prompted Lovable.dev how it would solve it.

The suggestion was simple: drop in a resend.com API key and call the function from inside the app. When I looked at the code, the only thing broken was a hardcoded key — easily fixed. What struck me was how natural the AI’s structure was. Logical, testable, and completely valid — just different from how I’d do it.

That moment flipped a switch. I realized I was spending time asking Lovable to do very complex things in very specific ways, just to see how far I could stretch the tool. And what I found was even more useful: it showed me the edge cases — the kinds of tasks I probably shouldn't be doing manually at all anymore.

4. Adoption Friction in Senior Teams

Over the past year, I’ve consistently seen that larger development teams—especially those with a DevOps or outsourced delivery model—show the highest resistance to AI tooling. In three different teams I spoke with, adoption hovered around 25%, with the remaining majority expressing doubt or avoiding use altogether. Even when companies went to great lengths to build local, HIPAA-compliant models and run them on top-tier Apple hardware, the investment didn’t lead to full-scale adoption. Demos worked. Performance gains were real. But most developers either didn’t engage (fully) or only did so in a minimal, half-hearted way.

These teams made it clear they weren’t interested in junior engineers using AI. They sought senior operators—engineers with sufficient experience to guide AI tools like Copilot or Cursor and extract 5–10 times the utility gains. The concern wasn’t just about output; it was about control, interpretation, and the ability to craft effective prompts. AI wasn’t viewed as a learning accelerator or junior assistant. It was seen as a high-leverage tool for high-context players.

Even in ideal technical environments, cultural and practical skepticism remained the blocker. Usage logs didn’t lie. Teams might have paid for AI access, but adoption stalled where belief and process didn’t evolve alongside the infrastructure.

5. Developers who don’t talk to customers

I have interviewed numerous developers. It has not been uncommon for developers to be super clear: I do not want to talk to customers. It's so common that job descriptions sometimes make that clear so that candidates do not opt out.

This is an issue now. Since not understanding the values and having them filtered through sales, product managers, and senior developers works against the goal of agility and getting ideas to market.

The outward expression of this challenge comes from CEOs when they ask if some of the development tasks can be moved to their product owner, so that the voice of the customer remains full strength and does not get diluted by the process.

I have a feeling that the days of a developer working alone in a back room might be coming to an end. AI relies on good prompts and good questions. Being disconnected from a customer is not ok.

Good News

Twice this month, I have seen the results of a senior developer who has gotten over their initial objections and applied AI tooling to existing projects. The results are excellent. It’s not hard to see why. If non-technical people who have never coded anything can build with AI, imagine what an experienced person can do!

If we get approval, we will highlight some of these projects. They are mind-blowing.

Looking for unbiased, fact-based news? Join 1440 today.

Join over 4 million Americans who start their day with 1440 – your daily digest for unbiased, fact-centric news. From politics to sports, we cover it all by analyzing over 100 sources. Our concise, 5-minute read lands in your inbox each morning at no cost. Experience news without the noise; let 1440 help you make up your own mind. Sign up now and invite your friends and family to be part of the informed.

Subscribe to 1440 today.