Software is becoming marketing

The layoffs are piling up fast enough that they've stopped being news individually. Block cut nearly 4,000 people. Atlassian dropped 1,600. Salesforce swapped out 4,000 support roles for AI. Snapchat shed 1,000. Even a mid-size Czech insurer, Direct, is cutting a third of its 1,200-person workforce, with its CEO flatly declaring that "the time of playing around with AI is over." That's not a Silicon Valley trend anymore, that's a global one.

But the layoffs are only the loud part. The quieter, weirder shift is what happens to software itself once anyone can vibecode a working app in an afternoon instead of paying $50 a month for someone else's. One writer, drawing on his own oddball background in math and marketing, makes the case that software is sliding from a math-like field, where outsiders can't even evaluate the work, into a marketing-like one, where everybody has an opinion on your landing page colors. Once the barrier to entry drops and everyone feels qualified to judge the output, respect for the craft collapses, and pay compresses toward what writers, designers, and marketers already deal with: wide pay gaps where only the proven best get paid well, and everyone else fights for scraps.

Then there's the choice-paralysis problem. When there were three project management tools, people picked one and stuck with it. When there are three hundred, and building a three-hundred-and-first takes an afternoon, nobody commits to anything. It's the same psychology as dating apps: infinite options make people pickier, not happier. Barry Schwartz called this the paradox of choice decades ago, and now it's coming for your CRM.

That abundance is expected to hollow out the middle of the SaaS market entirely. Companies charging modest monthly fees for solid-but-unremarkable tools lose their moat once a free AI clone is a weekend project away. What's left is a power-law split: a Cambrian explosion of tiny, disposable, personal tools built by individual freelancers and teachers for themselves, and a handful of giant aggregators that win purely on trust, distribution, and identity infrastructure, not on features.

The more interesting pivot might be from software to services. Sequoia has floated the idea that the next trillion-dollar company will be a software firm pretending to be a services firm, and the logic checks out: companies already spend roughly six dollars on services for every dollar spent on software. If AI can just close the books instead of selling accounting software, the money moves from tools to outcomes. That also reshapes what an