Does AI Decimate the Software Business or Grow it by 100X?

Jevon's Paradox

A recent Twitter exchange between Marc Andreessen and Ben Horowitz, found on Twitter, explored the surprising idea that building companies might become more expensive in the era of Artificial Intelligence (AI).

Andreessen argues against the common belief that AI will simply cut costs and boost efficiency. He introduces Jevon's paradox to suggest that while AI might decrease the cost of individual software components, the overall cost of building a software company could rise. Customers will demand ever-more-powerful features, leading to an innovation arms race as companies strive to meet these expectations.

Here's my perspective:

  • The Rise of Custom Software: I believe AI will primarily impact software by enabling the creation of more custom software for companies, rather than large, generic software packages. This trend already mirrors the assembly-line approach seen in automotive manufacturing, where open-source software components are combined to create custom solutions. This approach, coupled with low- and no-code tools, already reduces software costs. AI will likely accelerate this shift towards custom solutions.
  • The Software Building Challenge: Surprisingly, even software companies often struggle with internal software development. It's not that they lack the technical expertise, but rather the confluence of necessary skills:
    • Organizational Skills: Defining the internal need for software, securing budget approval, hiring and managing the right talent, and navigating stakeholder requirements.
    • Discovery and Design: Extracting product requirements, translating them into a design and a plan, and managing timelines and budgets.
    • Team Assembly and Management: Bringing together a team to deliver the final product.

The cobbler's children paradox may explain this software-company struggle. Design, engineering, and QA teams are laser-focused on maintaining existing revenue streams, so building new things is a challenge. Additionally, top talent often prioritizes working on established revenue generators, leaving new projects under-resourced.

  • AI and the Custom Software Boom: AI, along with readily available open-source components and improved tools, will further reduce the cost, time, and personnel needed to build custom software. This aligns with Jevon's Paradox – increased demand due to lower costs. However, the surge in demand will likely benefit the custom software outsourcing business.

For outsourcing companies to compete, they'll need to utilize AI and other tools to streamline their marketing, sales, and product discovery processes (building specifications and designs). This will mirror the compression of development and QA thanks to AI.


In essence, AI, coupled with existing trends, will likely compress the entire custom software development lifecycle, leading to an explosion in this market segment – a consequence of Jevon's Paradox at play.

To learn more about how AI can revolutionize software in Oil and Gas or how it can transform challenges into opportunities in banking, check the articles below:

Like what you see? Share it with your friends.
Mike Hogan

Mike Hogan

My team and I build amazing web & mobile apps for our companies and for our clients. With over $2B in value built among our various companies including an IPO and 3 acquisitions, we've turned company building into a science.

Leave a Reply

Your email address will not be published. Required fields are marked *