In the video, Cameron predicts that by 2028, AI will revolutionize software development through innovations like AI-driven UX testing, parallel codebases, and granular version control, enabling faster, more efficient, and user-focused coding processes. He also foresees a shift toward just-in-time software and integration-centric platforms, emphasizing seamless data connectivity over standalone applications.
In this video, Cameron, a co-founder of a successful software development company, shares his top five predictions for how coding and software development will evolve over the next three to five years, particularly influenced by advancements in AI-assisted programming. He acknowledges the widespread belief that AI might eventually replace human programmers but emphasizes that this transition will be gradual. Cameron draws from his experience using AI coding tools daily and observing emerging technologies like Cursor and Windsurf to forecast upcoming trends in the industry.
The first prediction focuses on “Agentic AI UX testing,” where AI bots simulate human users from specific demographics to test user experience flows on websites or applications. This approach could drastically reduce the time and resources currently required for traditional UX testing methods like user focus groups or A/B testing. These AI agents would mimic human behaviors, including distractions and errors, enabling rapid and scalable testing to optimize software interfaces more efficiently.
Cameron’s second prediction involves the emergence of “parallel agents” and parallel codebases. Instead of working on a single codebase with branches, multiple entire codebases could be developed simultaneously using different AI models or approaches. These parallel versions could then be tested—potentially using the agentic AI UX testing method—to determine which performs best for different user groups or scenarios. This would lead to a proliferation of code variations and faster innovation cycles in software development.
The third prediction is about “constant commits,” where version control systems evolve to track every single change in the codebase with extreme granularity. Given the rapid and extensive modifications AI-assisted coding can introduce, developers will need more detailed commit histories to understand, review, and manage changes effectively. This granular tracking would also enable better collaboration and allow developers to query the AI about the rationale behind specific code changes.
Finally, Cameron predicts a shift toward “just-in-time software” and a focus on integrations rather than standalone software products. Software will be rapidly generated for specific, immediate use cases and then discarded or replaced as needs change. Additionally, software ecosystems will become more interconnected through APIs, allowing users to aggregate and interact with data from multiple sources seamlessly. This integration-centric approach will make individual software less important than the platforms and data connections they enable, transforming how users access and utilize digital tools.