There’s an A16Z podcast episode going around about how Product Managers, Engineers, and Designers all think they can do each other’s jobs now that AI is here.
The truth is more interesting: if you’re already good at two of the three, AI can close the gap on the third. And that changes everything about what a solo founder can build.
Every startup textbook tells you the same thing: you need three core competencies to build a product.
Product Management — Understanding users, defining problems, prioritizing ruthlessly
Engineering — Building the thing
Design — Making it usable and beautiful
The conventional wisdom is that no one person can do all three well. So you find co-founders. You split equity. You spend months aligning on vision.
This made sense when each skill required years of specialized training and the tools were primitive.
It makes less sense now.
AI doesn’t replace any of these skills. What it does is lower the activation energy for adjacent work.
If you’re a PM who can code a little: AI turns “a little” into “enough to ship.” Claude and Copilot don’t write your architecture, but they write the boilerplate, catch your bugs, and let you focus on the 20% that matters.
If you’re an engineer who struggles with design: AI tools generate decent UI starting points. You still need taste to refine it, but you’re no longer staring at a blank Figma file.
If you’re a designer who can’t code: AI bridges the gap between mockup and implementation. The “handoff” problem is dissolving.
The pattern: AI compresses the distance between “I understand this conceptually” and “I can actually execute it.”
I spent 20 years in product at eBay, Ariba, and several smaller companies. I’ve shipped features to billions of users. Product is my native language.
I can code — not at a senior engineer level, but enough to be dangerous. I built the Voice Legacy backend, the AI conversation system, the whole thing. It took longer than it would take a “real” engineer. But it shipped.
Design is my weakest link. But AI-generated UI components, combined with component libraries and iteration, get me 80% of the way there. The last 20% I hire for eventually.
The math: Two strong skills + AI assistance on the third = viable solo founder.
I could have spent six months finding co-founders. Very few good engineers want to execute on a ‘business guy’s’ vision, because as a former engineer that just sounds like a nightmare.
Instead, I spent those six months building.
Voice Legacy exists because I needed it. My dad died, and I only captured 10% of his stories. I didn’t want to wait for a co-founder to feel the same pain.
AI let me move fast. Not by replacing my judgment, but by handling the parts where my judgment wasn’t the bottleneck.
This only works if you have genuine depth in at least two areas. AI amplifies competence; it doesn’t create it from nothing.
If you’re a PM who has never written a line of code, AI won’t make you a developer. If you’re an engineer who has never talked to a customer, AI won’t give you product sense.
But if you’ve spent years building adjacent skills — coding side projects as a PM, developing product intuition as an engineer, learning enough design to have opinions — AI is the multiplier you’ve been waiting for.
Old model: Find co-founders, split equity, align on vision, then build.
New model: If you’re good at 2/3, AI closes the gap. Build first, hire specialists later when you have traction and revenue.
I’m not saying co-founders are obsolete. For some products and some founders, they’re still the right call.
But the bar for “you need a co-founder” is higher now. The solo founder path is more viable than ever — if you’ve put in the work to be genuinely good at multiple things.
I’m Henry, founder of Voice Legacy. We help families preserve stories through AI-guided conversation. If you have aging parents whose stories you want to capture, I’d love to show you how it works.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.