Code Never Mattered.

The idea that every company will now build their own SaaS for their own operations is ludicrous for a very simple reason: if we’ve learned anything from the past two years, it’s that customers are lazy, and that’s okay.

This has never been about skill set. Code has always been cheap. You’ve always been able to code a DoorDash alternative or a Bloomberg alternative. When the source code of companies like Microsoft has leaked, and it has, frequently, the real harm was never people copying and distributing it for free. It was security vulnerabilities. Something that, if anything, will only be heightened with the advent of code generation.

Code doesn’t matter because it’s always been about traditional business moats.

If you don’t believe me, look at the numbers. In fiscal year 2025, Salesforce generated $37.9 billion in revenue. They spent approximately $5.5 billion on R&D, roughly 14% of revenue. They spent more than double that, around $13 billion, on sales and marketing. That’s a company that understands what actually drives growth. For every dollar Salesforce puts into writing code, they put over two dollars into distribution, relationships, and market presence. The code is table stakes. The business is the product.

If you think code matters, that tells me you’ve never worked at a company of more than a hundred employees. Past a certain size, software companies stop being software companies. They’re just companies. They’re just businesses. And businesses are about two things and two things only: revenue and profit.

The way you increase revenue is not by making your code better or even your NPS better. It’s distribution. It’s lock-in value. It’s contract structure. It’s pricing structure. It’s knowing extremely well where you sit in your customer’s P&L.

It’s never been about the code. And now it matters even less.

What does this mean for SaaS?

SaaS companies that have made it past a certain critical size are no longer software companies. They’re just companies. Code doesn’t matter. Only functionality and outcomes do.

Whenever you hear criticism on X about how many employees DocuSign has, remember: most of those employees are not writing code every day. They’re there to sell. They’re there to make their own salary and then some.

But here’s what has changed: the bar has increased, ever so slightly.

When writing code is this easy, there is no excuse not to ship improvements that increase your customers’ NPS. There is no reason to have shitty user experience.

Gone are the days where companies like Salesforce or HubSpot can afford subpar UX and retain customers, not because their customers are going to build an in-house competitor, but because AI-native competitors with better user experience are going to eat their lunch.

There comes a time in the lifecycle of a software company where you need re-founders. Re-founders are people who carry and expand the vision of the company in a naive, necessary way. They don’t think about the company in terms of dollars, but in terms of user experience and qualitative attributes. Think of the leader who walks into a $200M ARR company and asks “why does this workflow take nine clicks?” instead of “how do we get to $300M?” The second question answers itself when you fix the first.

The days where Salesforce can maintain market influence by sheer inertia, by giving out free tickets to the Warriors, those days are numbered. The bar has increased. And that’s great.

What does this mean for the future?

Five years from now, “software company” won’t be a category. We’ll stop talking about software companies entirely. Companies have always just been companies. Companies are about the P&L and the balance sheet.

Now, don’t mistake my directness for vapidness. I truly believe that good UX is better for society, and that there is intrinsic value in the qualitative attributes of a well-run company with a great customer experience.

But when you read the next X post saying Perplexity killed Bloomberg or Claude killed DoorDash, it’s for the wrong reasons. It’s never been about replicating. Replicating is easy. You’ve always been able to hire talent in an emerging market and spin out copycats. If anything, HBO’s Silicon Valley explicitly made fun of this years ago.

The hard part has always been packaging, marketing, sales, and distribution. That’s it. That’s all that matters to generate shareholder value. If you can pair that with an incredible vision, the vision is the conduit for all of it to happen. But ultimately, a company needs to deliver results in the form of dollars.

So no, Perplexity did not kill Bloomberg. Bloomberg will be just fine.

Bloomberg generates an estimated $15 billion a year in revenue. Over 325,000 subscribers pay upwards of $24,000 per seat annually, locked into two-year contracts. Dozens of competitors have tried to undercut them over the past four decades. Not one has made a dent. Bloomberg’s market share has held at roughly a third of the global financial data market for as long as anyone can remember. The code was never the moat. The contracts, the data relationships, the network effects of 325,000 professionals messaging each other on the same platform, that’s the moat.

When everyone can make their own Bloomberg, the actual Bloomberg becomes much more valuable. They already have the momentum. They already have the inertia, the contracts, the distribution, the credibility.

The market is about to find out that the concept of a “software company” is dead.

Software companies are no longer going to be about software. They’re just going to be companies.

Like they always should have been.