> ## Content Index > Fetch the complete content index at: https://www.ceo.com/llms.txt > Use this file to discover other available public pages before exploring further. # Software is no longer something you buy - URL: https://www.ceo.com/blog/software-is-no-longer-something-you-buy/ - Published: 2026-01-16T01:22:00.000Z - Updated: 2026-08-14T01:26:14.000Z - Description: For decades, technical skill was the bottleneck. Remove it, and the plumber who understands plumbing beats the developer who doesn't. - Author: CEO.com - Tags: #blog The most interesting thing happening with AI isn't which startup will capture the market. It's that the market itself might be disappearing. The revolution won't come from a new company selling software to everyone. It will come when everyone can build the software they need without buying it from anyone at all. Claude Code, Anthropic's AI coding tool, has gone viral. Developers are calling it the best coding assistant available. [Anthropic announced](https://www.anthropic.com/news/anthropic-acquires-bun-as-claude-code-reaches-usd1b-milestone?ref=ceo.com) it has hit $1 billion in annual recurring revenue. But more interesting than what it does for programmers is what it signals for everyone else. [Axios](https://www.axios.com/2026/01/07/anthropics-claude-code-vibe-coding?ref=ceo.com) put it plainly: "A mass market, easy-to-use coding tool is a game changer for the 99% of people who, until now, had to rely on other people's software to do anything from building custom data tools to automating repetitive work to building apps or websites." Then Anthropic released [Claude Cowork](https://venturebeat.com/technology/anthropic-launches-cowork-a-claude-desktop-agent-that-works-in-your-files-no?ref=ceo.com), extending that power to non-technical users. No terminal. No command line. You give Claude access to a folder on your computer, describe what you want, and it reads, edits, or creates files on your behalf. Organize a cluttered downloads folder. Turn receipt screenshots into expense spreadsheets. Draft reports from scattered notes across multiple documents. The most remarkable detail: a team of four built Cowork in roughly 10 days using Claude Code itself. AI built the tool that enables non-developers to do complex work in less than two weeks. In February 2025, Andrej Karpathy, a co-founder of OpenAI and former AI leader at Tesla, coined a term for something people were already experiencing: "vibe coding." He described it as fully giving in to the vibes, embracing exponentials, and forgetting that the code even exists. You describe what you want. The AI builds it. You never look at the underlying code. Dan Shipper, CEO of the AI subscription service Every, called Claude Code an "[infinite vibe coding machine](https://www.axios.com/2026/01/07/anthropics-claude-code-vibe-coding?ref=ceo.com)." Kevin Roose at the New York Times experimented with it and came away calling his creations ["software for one"](https://www.nytimes.com/2025/02/27/technology/personaltech/vibecoding-ai-software-programming.html?ref=ceo.com): personalized tools designed for specific individual needs. An app to analyze his fridge contents and suggest what to pack for lunch. Not production-grade software meant to scale. Just a small tool that solved a specific problem for exactly one person. This sounds trivial until you think about what it means. Gartner predicts that low-code development will account for [more than 70% of all application development activity this year](https://www.gartner.com/en/newsroom/press-releases/2021-11-10-gartner-says-cloud-will-be-the-centerpiece-of-new-digital-experiences?ref=ceo.com), up from 20% in 2020\. Y Combinator reported that 25% of startups in its Winter 2025 batch had codebases that were 95% AI-generated. McKinsey estimates generative AI will reach $175 billion to $250 billion by 2027. The consulting firms are paying attention. [Bain's Technology Report 2025](https://www.bain.com/insights/will-agentic-ai-disrupt-saas-technology-report-2025/?ref=ceo.com) frames this as a disruption that will "commoditize" some SaaS markets. [Harvard Business Review](https://hbr.org/2025/05/how-gen-ai-could-disrupt-saas-and-change-the-companies-that-use-it?ref=ceo.com) published a piece arguing that generative AI will transform traditional workflow systems into dynamic, goal-oriented environments. The language is careful and hedged, as consulting language tends to be, but the underlying message is clear: software as we've known it is changing. But here's what interests me more than the disruption of SaaS companies: the empowerment of people who were never in the software conversation at all. Think about all the small problems that exist in businesses across the country. The yoga studio that needs a booking system with specific features no vendor offers. The HVAC company that wants to integrate their scheduling with inventory in a way that matches their actual workflow. The restaurant owner who needs something better than the generic point-of-sale system that doesn't quite fit how they operate. These problems are too small to justify a startup. No venture capitalist is going to fund a company to build scheduling software for HVAC companies in Utah. The market isn't big enough. So these businesses either buy something generic that doesn't quite work, duct-tape multiple tools together, or live with the friction. What happens when the owner of that HVAC company can describe exactly what they need and have AI build it for them? Anthropic noticed something unexpected with Claude Code: users were deploying the developer tool for tasks unrelated to programming. [Boris Cherny](https://venturebeat.com/technology/the-creator-of-claude-code-just-revealed-his-workflow-and-developers-are?ref=ceo.com), the head of Claude Code, listed what people were using it for: vacation research, building slide decks, cleaning up email, canceling subscriptions, recovering wedding photos from a hard drive, monitoring plant growth, controlling an oven. These use cases are diverse and surprising, he wrote, because "the underlying Claude Agent is the best agent." This inverts the current hierarchy in ways that matter. For decades, technical skill has been the bottleneck. You had an idea, but you needed engineers to build it. You understood your industry, but you needed developers to translate that understanding into software — the people who controlled technical resources controlled what got built. If AI removes that bottleneck, domain expertise becomes the scarce resource. The plumber who deeply understands the plumbing business is more valuable than the developer who can code but doesn't understand plumbing. The constraint shifts from "can we build it?" to "do we understand what should be built?" Of course, the skeptics make fair points. Gary Marcus, the cognitive scientist, has [criticized the hype](https://garymarcus.substack.com/p/decoding-and-debunking-hard-forks), noting that it's often easy to achieve 80% accuracy but extremely difficult to reach 20%. Scott H. Young, writing about his experiments with vibe coding, observed that even though he wasn't writing the code himself, he still needed significant [conceptual knowledge](https://www.scotthyoung.com/blog/2025/11/12/vibe-coding-future-work/?ref=ceo.com) to get good results. There are real concerns about security, maintainability, and the risks of people building things they don't fully understand. A Hacker News commenter made an analogy I found useful: almost anyone can do basic DIY work around the house, but when it comes to serious work, you still need contractors. Vibe coding might be similar. It expands what individuals can do themselves while actually increasing demand for professional help when things get complex. Which raises another question: What will VCs invest in if anyone can build their own software? The venture model depends on markets being large enough to justify investment. Build something that can scale to millions of users, capture a significant market, and generate returns that justify the risk. But what if the future is millions of tiny markets, each served by bespoke solutions built by the people who understand them best? The SaaS model depends on customers not being able to build their own solutions. That assumption is weakening. When software becomes something you describe rather than something you buy, the entire investment thesis behind most software companies starts to look fragile. Maybe VCs will shift toward infrastructure: the foundational AI models, the cloud computing that powers them, the chips that make it all possible. Maybe they'll focus on tools that help people build and manage their own software. Maybe they'll find entirely new categories that don't exist yet. The honest answer is that nobody knows. The playbook that worked for the last two decades was built on assumptions that are changing faster than most investors want to admit. Software is no longer something you buy; it's something you describe. That changes who gets to build things. That changes the world.