What Makes a Business Last When Software Is Cheap

Here's a thing I've been sitting with: Y Combinator — the accelerator that launched Airbnb, Dropbox, and about seven thousand other companies — recently held a session in Paris where its partners fielded questions from aspiring founders about how to build in the AI era. The whole thing was filmed and posted online, and it's mostly aimed at people who are trying to raise venture capital and conquer global markets.

So why am I writing about it here?

Because buried inside about 47 minutes of startup-world vernacular is something that matters to anyone who owns a business — a bookkeeping firm, a catering operation, a landscaping company with two trucks and a prayer. The partners kept circling back to a question that sounds like it's only about startups but is actually about everyone who builds anything: what is the hard part of your business, and is it still hard enough to matter?

The question behind the question

The way YC partner Tom Blomfield put it, stripped of the tech-world framing: ten years ago, building software was itself the hard part of building a software company. Writing code took time, expertise, and real money. That difficulty created protection — competitors couldn't easily copy you because copying you was expensive.

That protection is mostly gone now. AI tools have made software cheap enough to build that the code itself is no longer the thing worth defending.

"You have to have something that's hard," Blomfield said. "And that might in some cases just be brutally hard sales into some industries. Others it might be a regulatory barrier... most hardware in and of itself is just like hard. The atoms add the hard bit."

Now, if your business was never about software in the first place — if you're a plumber, a restaurant owner, a freight broker — you might read that and think: okay, not my problem. My whole business is the hard part. Nobody's automating a tile installation or a birthday cake consultation.

Fair. But the flip side of that observation cuts a different way for Main Street: AI tools are about to (or already have started to) change what your competitors can do cheaply. The bookkeeper who used to take three days to produce a report can now produce it in three hours. The graphic designer your competitor paid $400 now costs $40. The "hard part" of a lot of service businesses wasn't their expertise — it was the friction of accessing that expertise, the time cost of production. Some of that friction is dissolving.

Which means the question Blomfield posed to a room of Paris founders is actually a useful question for anyone with a business: what is the hard part of what you do, and is it durable? The answer might be your relationships. It might be your physical presence in a community. It might be trust built over fifteen years of showing up. Those things don't get cheaper when AI improves. That's the part worth knowing.

The partner question

The session also spent real time on something that I find endlessly relevant regardless of your industry: who you go into business with, and why.

The question in the room was about co-founders, framed around Sam Altman's much-discussed prediction — reported by TechCrunch — that AI could soon enable the first one-person billion-dollar company. The partners acknowledged the theoretical possibility and then, fairly gently, said they thought it was a bad idea anyway.

Their reasoning wasn't about capability. It was about the hardest days.

Blomfield described funding a strong solo founder that batch who had a crisis of confidence six weeks in. "He's like, I'm not sure this is working, is this the right direction? I'm like, dude — this is what a co-founder is for. It's someone to pick you out of the gutter when you're feeling down, or when you're hyperactive and manic, they pull you back down to earth and they modulate you."

That's not a startup observation. That's a business observation. The number of people I've talked to over the years who went into business alone and hit a wall — not a strategy wall, a human wall — is significant. The ones who made it through usually had somebody: a spouse who understood the stakes, a trusted employee, a peer in the same industry they called every few weeks. Formal co-founder or not, the function Blomfield is describing — the person who witnesses what you're going through and doesn't let you spiral — is not optional for most people.

What the YC partners said about choosing that person is where I'd push back a little. Their advice was to prioritize character over complementary skills. Find the smartest, most determined, highest-integrity person you've ever worked with, and don't worry too much about whether their skills fill your gaps. "The business stuff is actually very easy to learn," Blomfield said.

I'd love to believe that. But I've watched enough people fail at the relationship-capital side of running a business — the networking, the community trust, the reading-the-room that keeps a loyal customer from leaving — to know that dismissing it as "easy to learn" is a perspective shaped by a particular kind of founder. Technical people who went on to build very large companies. The business stuff might be learnable for someone with the time, runway, and support to learn it. For someone bootstrapping a catering company while working a day job, that skill gap can be the whole ballgame. Choose your partner knowing that.

The one thing not to automate

The most practically useful exchange in the whole session was brief. Someone asked the partners: if you were starting over, what's the last thing you would automate?

Blomfield didn't hesitate: "Talking to customers."

"It was the thing that meant I was the person closing the loop and had all the context on what it is that we should build. I never wanted to stop having that information."

Another partner added: don't delegate the conversation with your co-founder, either. The companies that fail during a YC batch often fail because the two founders stopped having real conversations with each other.

Both of those things are true at every scale of business. The owner who stops talking to customers — who delegates it fully to staff, who starts managing by report rather than by conversation — loses the thread of what's actually happening in their business before they know they've lost it. You can automate the follow-up email. You cannot automate the thing you learn when a longtime customer says, "Actually, I've been meaning to talk to you about something." That information doesn't live anywhere else.

What the pivot question actually reveals

There's a useful concept buried in the session's conversation about pivoting — about when to change direction on something that isn't working. The partners' framing, stripped down: most pivots happen not because the fundamental idea was wrong, but because the founder got sad and rejection-fatigued and started looking for an exit from the discomfort.

A real pivot, they said, happens when your customers pull it out of you — when you're deep enough in the work that the better version of the idea reveals itself. One partner described GoCardless, which according to Grokipedia began by attempting to collect payments for sports clubs before its users made clear they wanted the same capability for their businesses. The pivot wasn't a strategic retreat. It was the market showing you the door.

The more common version is just burnout dressed up as strategy. The grass-is-greener pivot. The partners were blunt about it: the grass is not greener. You'll just find a new set of problems you didn't know existed.

That applies whether you're a founder or an established small business owner thinking about adding a new service line, or dropping an old one, or chasing a trend because the current model feels slow. The useful question is not is there a better idea over there? It's have I actually exhausted this one, or am I just tired?

---

The YC Paris session was aimed at founders who want to build companies worth billions. Most of the people reading this are not those founders. But the session kept surfacing the same durable truths that every business rests on: the hard parts worth protecting, the people worth building with, and the conversations you should never stop having yourself.

That part doesn't change with the technology.

---

Dorothy "Dot" Williams covers small business and Main Street economics for Buzzrag.