The AI race is in full sprint, and it has created a new reality for builders: as soon as cutting-edge breakthroughs are discovered, they’re shipped. This has leveled the playing field in a way I don’t think we talk about enough.
For maybe the first time ever, product builders, developers, designers, and founders all have access to the same ridiculously powerful technology at the same time. I’ve been talking to a lot of founders lately, both current and aspiring, and I can’t stress this enough: now is a truly great time to be building.
But to seize this moment, we have to be willing to question the way things have always been done. AI enables new capabilities, but more importantly, it demands entirely new workflows. This isn't just about making your old process faster; it's about fundamentally rethinking the process itself. And that's why it's so important to go back to first principles - which, I've learned, can be deeply uncomfortable.
Leaning Into the Discomfort of Innovation
I had a moment in a meeting this week that’s been rattling around in my head. A colleague made a quick, off-the-cuff comment about a product area I'm trying to redefine. Their feedback was logical and rooted in what works today, and it directly challenged my more ambitious, 10x thinking. It threw me off.
This morning, I realized why it bothered me: when you're deep in 10x thinking, that kind of feedback can feel like a gravitational pull back to the safety of the status quo.
We often confuse "uncomfortable" with "wrong." But in true innovation, discomfort is often a sign that you're pushing into genuinely new territory. It's a signal that you're leaving the well-trodden path. My takeaway wasn't to retreat, but to lean in and understand the "why" behind their perspective (and the “why” behind my reaction!), using it to sharpen my own conviction. It’s a "discomfort filter"—a way to separate valid challenges from fear of the unknown.
Automating the Old vs. Inventing the New
This brings us to a classic question when applying AI: do you simply make an old process faster, or do you invent a new one? My take is that starting with the former is a great way to get going, but you shouldn't stop there.
Take my resume review workflow. The old way was hours of manual sifting. My first thought was to apply AI to simply speed that up. I decided to build an Opal that could review resumes for me. But to do that, I first had to create a detailed rubric of the qualities I actually look for in a candidate. The act of building the AI tool forced me to crystallize my own thinking.
The first results were... too perfect. The AI, eager to please, gave almost every resume a glowing score, pattern-matching the rubric without real discernment. So, I iterated. I fed the rubric and a few manually graded resumes back into Gemini, explaining my reasoning for each, and asked it to create a new meta-prompt for itself - one that could critically and fairly evaluate future resumes.
And it worked. My new workflow is to run resumes through the Opal, and then I do a final gut check on the AI's prioritized list. This is a huge win: it has dramatically sped up my work, and it forced me to deeply codify my own evaluation metrics, making my entire hiring process better. This is a perfect example of AI making an old process more efficient.
But it's still an optimization. The truly transformative power of AI comes when it enables you to invent an entirely new process.
This brings me to a more profound shift: my writing workflow. For years, the process was standard: sit down, stare at a blank page, and carefully craft each sentence. My new, AI-native workflow is completely different.
It started with a significant upfront investment: I spent a long time perfecting a "master prompt," a detailed System Instruction for Gemini that encodes my unique style, tone, audience, and even specific editing rules like "make sure every word counts."
With that foundation in place, my process has been transformed. I now do a lot of my writing as voice memos while my husband drives us to work. I can capture ideas on the go, sometimes even recording a thought, pausing to talk it through with him to solidify my thinking, and then resuming the recording.
From there, I can feed that raw, unstructured input into AI Studio and get back a coherent first draft that already sounds like me. I'm still deeply involved - I read over every word, edit heavily, and usually have at least one other person proofread my posts. But the change is profound.
This new workflow isn't just about speed, although it is much faster. The real magic is that it helps me get more of my ideas out of my head quickly, which in turn helps me to think and connect more dots. And I'm not the only one discovering this; a coworker recently told me she now includes transcripts of her spoken thoughts as additional context for her team's performance reviews, finding it helps her capture more nuanced ideas and improve the quality of her feedback. It's a fundamental shift from capturing to crafting. This isn't just automating my old writing process; it's an entirely new way of creating.
The Myth of the "Secret Weapon"
This brings me to another conversation that has stuck with me, this one with a founder friend building something new in a domain ripe for disruption. He was getting distracted by a nagging thought: "What am I missing?" He kept asking what secret tool a big studio might have, or what special technology a famous director was privy to that he wasn't.
Here’s the thing: there is no secret weapon you don't know about.
The advantage of large incumbents is also their greatest weakness: they are often anchored to established ways of working. You, the disruptor, have the freedom to recognize that it's not just the playing field that's changing - it's the sport itself. Instead of asking "What am I missing?", the only question that matters is "What can I build?"
Questions to Kickstart First-Principles Thinking
My perspective, especially for disruptors, is clear: don't get bogged down trying to recreate the old way of doing things. Go back to first principles and rethink the entire problem space. To get started, here are a few questions I find helpful:
The Fundamental Need Question: "What is the core human problem this old process is trying to solve, and is there a more direct path to that outcome with today's AI?"
The Future-Proofing Question: "If the core AI technology I'm using becomes 10x better and free next year, is my product still valuable? Where does my true defensibility lie?"
The "Magic Wand" Question: "What 'human-in-the-loop' step in my current idea exists only because of a technical limitation, and what would I build if that limitation disappeared tomorrow?"
The new sport is being played in a world where the value of a user's time is skyrocketing. The bar for what people expect from a product is only going to get higher. The companies that will win are the ones pushing the boundaries and rethinking what's possible.
So, think big. Think different. Then get out there and build.
"Discomfort as a filter" is a powerful way to recognize when we’re actually pushing into new territory rather than slipping back into the status quo. I loved the distinction between “automating the old” and “inventing the new”; the writing workflow example makes that shift tangible. And the reminder that there’s no secret weapon, only the willingness to rethink from first principles, is inspiring and practical at the same time.