My 2025 AI-First Pledge: Building Better Products Through Daily Use
Unlocking the next generation of product innovation requires more than just understanding AI; it demands immersion.
As 2025 kicks off, I've been doing a lot of thinking about my goals for the year. And there's one that stands out: I'm committing to using AI every single day. Why? Three key reasons:
Sharpen my product instincts.
"10x myself" and boost productivity.
Unlock capabilities I never thought possible.
It might seem a little forced at first – this whole "AI-first" approach. I expect it to feel a bit weird, maybe even unnatural, initially. But I'm convinced it's going to unlock some incredible insights into building better, more innovative products and experiences.
Let's dive a little deeper into that first point: forming better product opinions. I've long been a proponent of "tech-first" product development. In this era of rapidly advancing AI, I think you need to double down on this. You can't rely on users to tell you what to build because, frankly, they don't know what's possible. Most (almost all) haven't seen behind the curtain. To build truly interesting, groundbreaking products, we need a "tech-first" approach that's deeply rooted in our own understanding of the technology. That's where daily AI usage comes in. By immersing myself in AI's capabilities, I'll gain a far more intuitive understanding of what's possible, allowing me to identify opportunities to use AI to build experiences that address needs and desires users don't even know they have yet. As Steve Jobs famously said, "People don't know what they want until you show it to them." And if you’re going to show them something truly revolutionary, you first need to know what's possible yourself. Daily AI usage is the key to unlocking that knowledge.
The other thing I've come to realize is that building AI-first products requires a fundamental shift in thinking. The workflows are different. AI-native products demand a different approach. They aren't just about bolting AI features onto existing frameworks; they're about reimagining processes from the ground up.
Take writing a story, for example. A while back, I wrote a post called "Destined for Average: Why This AI Story Experiment Fell Flat." It explored a simple premise: a published author and ChatGPT were given the same prompt and challenged to write a short essay. Readers then had to try to tell them apart. The result? AI's output was, well, average. The prompt itself was remarkably basic: "Write a thousand-word short story in the style of Curtis Sittenfeld that includes these elements: lust, kissing, flip-flops, regret, and middle age."
Now, as someone who's been using AI to help with writing, I know that's not the most effective way to leverage its power. If I were using AI to write a story, I'd break the process down into manageable steps. Let's use my childhood hero, Marvin the Magician (thanks, Mom, for those amazing bedtime stories!), as an example. To craft a compelling Marvin story, I'd do something like this:
Character Profile: Use AI to help me create a detailed "character profile" of Marvin, fleshing out his personality, quirks, and backstory.
Plot Brainstorm: Use AI to brainstorm a compelling plot for a story featuring Marvin.
Chapter Outline: Prompt AI to break down that plot into a series of chapters, defining the high-level concept for each.
Chapter Generation: Prompt AI to write one chapter at a time, refining and iterating as I go.
And within that flow, I'd also use AI to critique the output. It could critique each chapter, at the end, or maybe even both! I'd probably even use AI to help me develop a rubric for critiquing the story.
Then, after all that, I might just scrap the whole workflow and try something completely different – informed, of course, by what I learned from the first attempt. That's "building to learn" in its purest form. And it all starts with using AI. So, when I say I want to use AI daily, I mean it. I want to experiment, iterate, and improve my understanding of these new workflows.
This commitment to daily AI usage isn't just about staying informed; it's about staying ahead. Because one thing's for certain in the world of AI as we kick off 2025: the models are only going to get better, faster. This constant evolution has two major implications for how I'm thinking about product development:
Ride the AI Tailwinds: We need to focus on products that leverage improving AI capabilities as a powerful force for growth. The better the models get, the stronger your product becomes.
Anticipate, Don't React: We should be designing products that assume model capabilities beyond what's currently available. This ensures your product feels cutting-edge at launch, not outdated.
Using AI daily will give me a better sense for the AI tailwinds that are worth riding. Multimodal is a big one, but getting more granular than that is crucial. And that level of insight only comes from hands-on experience.
And as for anticipating where AI is headed – well, that's another reason to use it every day. The pace of change is mind-boggling. We often think of progress as linear, like this:
But AI is accelerating so fast that it's hard to appreciate unless you're constantly experimenting. It's more like this:
So, that's my plan for 2025: embrace the "AI-first" mindset, use it daily, and see where it takes me. I'm excited to see what unfolds - and if you’re interested in following along how I’m using AI daily, you can follow me on X.