AI's primary impact on tech jobs is not direct replacement, but rather the diversion of corporate budgets from headcount to GPU infrastructure.
The most significant unsolved problem in software is not writing code, which AI does well, but the complex process of getting software into production.
The current cost of AI development tools for users is artificially low due to massive, unsustainable subsidization by platform providers.
AI will ultimately democratize software creation to the point that "everybody is a developer at some level," fundamentally altering the nature of the profession.
Instead of diminishing skills, AI coding assistants are augmenting developers by exposing them to new libraries, languages, and concepts they otherwise would not have encountered.
December (Prior Year)
Holland notes a "skyrocketing" use of the Claude model for coding and a near-doubling of open issues in the VS Code repository, indicating a surge in AI-driven development activity and complexity leading into the new year.
January 5th
In a blog post, Holland makes the bold prediction that the AI model Opus 4.5 would "change everything," setting high expectations for its impact on software development.
Post-January 5th
Holland validates his prediction through practical application, successfully building a native Windows tool with Opus 4.5 and a complete iOS application using the Gemini model in a single afternoon.
Present
Holland announces that GitHub is operationalizing advanced AI capabilities for enterprise use, with the GitHub Copilot CLI becoming generally available and new agentic workflows with memory and research features being introduced.
▶The Democratization and Commoditization of Software
Holland argues that AI drastically lowers the barrier to entry for software development, enabling individuals to build custom applications in hours or days that previously required teams or expensive SaaS products. This trend empowers more people to become developers at some level, but also commoditizes software, creating a significant risk for incumbent companies whose products can be easily re-implemented.
Investors should re-evaluate the moats of SaaS companies, as technological barriers to entry are eroding, shifting the competitive advantage towards distribution, brand, and data.
▶The Shifting Economics of the Tech IndustryMay 2026
According to Holland, AI is causing a major capital reallocation in the tech sector. Companies are diverting budgets from hiring (headcount) to investing in computational infrastructure (GPUs). This is compounded by the current, unsustainable business model where AI platforms heavily subsidize user costs, a practice Holland predicts will not last.
The end of heavy subsidization for AI tools could trigger a market consolidation and create significant cost pressures for companies that have become reliant on cheap AI-powered development.
▶AI as Augmentation, Not ReplacementMay 2026
Holland's perspective is that AI excels at the "easy part" of software development: writing code. The truly difficult challenges—architecting, testing, and shipping production-ready software—remain human-centric problems. In this view, AI acts as a powerful assistant that expands a developer's knowledge and handles tedious tasks, rather than making the developer's role obsolete.
The value proposition for skilled developers is shifting from pure coding proficiency to higher-level skills in system design, deployment strategy, and ensuring the reliability of AI-generated systems.
▶The Evolution of AI-Native Development WorkflowsMay 2026
Holland details the emergence of sophisticated, AI-centric development practices. This includes orchestrating multiple specialized AI models within a single task, using new agentic features in tools like GitHub Copilot for research and deployment, and grappling with the inadequacy of old testing paradigms for verifying AI-generated code.
A new generation of development tools focused on AI orchestration, verification, and agent management is likely to emerge, representing a significant investment opportunity.