The AIDEs Framework: A Unified Theory for AI in Software Development
JetBrains' framework maps the evolution and future trajectory of AI-powered developer tools.
1AI's entrenched role in development
AI has become ubiquitous in developers' workflows, with adoption extending even to traditionally slow-moving large enterprises. The technology is no longer experimental but a permanent fixture in software creation. Agentic coding, where developers delegate coding tasks to AI agents, is reshaping the traditional development loop from 'code → validate → fix' to 'plan → execute → review' in agentic chat interfaces.
2Claude Code's market dominance
Anthropic's Claude Code has emerged as the leading AI coding tool, used by approximately 8.5 million professionals and generating $7B in annual revenue. Its success has sparked a wave of IDE-agnostic CLI coding agents from competitors like OpenAI and Google, along with numerous niche players. The tool's broad acceptance across all categories demonstrates the maturity of AI-assisted development.
3Cloud agents gaining traction
AI agents are increasingly moving to cloud-based solutions like Devin, marking a significant shift in development practices. With more capable models and developers better understanding AI's capabilities, there's growing willingness to delegate simple, low-effort tasks to autonomous cloud agents. These agents can handle routine activities such as fixing linting errors during CI runs, though they still face substantial limitations.
4Uneven AI adoption across SDLC
While AI use is growing steadily in development, its application remains concentrated in just two areas: brainstorming and coding. Large portions of the software development lifecycle remain untouched by AI, indicating significant room for expansion. This uneven adoption pattern suggests future growth opportunities beyond the current focus areas.
5Evolution of AI in development tools
The progression of AI in development tools has followed a clear trajectory: from single-line code completion to multiline generation, then IDE-integrated AI assistants, followed by agentic development environments, and now cloud-based autonomous agents. This evolution reflects increasing delegation of development tasks to AI systems, with each stage building on previous capabilities.
6Core framework definitions
The AIDEs Framework introduces key concepts: Artificial Intelligent Systems (AIS) that demonstrate intelligence in helping users achieve goals, Artificial Intelligent Development Environments (AIDEs) specifically for software creation, and the principal-agent relationship borrowed from economics. These definitions provide the foundation for understanding AI's role in development workflows.
7Software creation vs development
The framework deliberately uses 'software creation' instead of 'development' to emphasize that the process involves more than just coding. It encompasses multiple roles and delegation relationships, from product teams to engineers to UX designers. AI introduces new actors into these existing structures without fundamentally changing the underlying organizational dynamics.
8Four foundational assumptions
The framework rests on key premises: 1) Continued growth in software demand, 2) Gradual delegation to AIS as the primary market driver, 3) Humans retaining specification and oversight roles, and 4) Decreasing personal immersion in delegated activities. These assumptions shape the framework's predictions about AI's evolving role in development.
9The delegation imperative
Historical trends from division of labor to automation show humans consistently delegate routine work. In development, higher-level tasks are delegated between humans, while lower-level tasks go to automated systems. As AIS capabilities grow, they'll become essential in this delegation chain, with the level of delegation to AIS serving as the key metric of their real impact.
10Future developer roles
AI won't eliminate developer jobs but will transform them. The emergence of 'junior architects' may require rethinking computer science education as high-level specification and oversight, traditionally senior roles, become more accessible. Developers will focus more on requirements formulation while delegating lower-value activities like code authoring to AI systems.
11Three-dimensional framework
The AIDEs Framework analyzes AI in development through three dimensions: stages of the software creation process (35 activities across 5 groups), levels of delegation, and organizational context. This multidimensional approach provides a comprehensive model for understanding current implementations and predicting future developments in AI-powered tools.
12Context management challenge
As delegation increases, developers face 'loss of immersion' in specific activities, similar to current human principal-agent relationships. Solving this through better context abstraction and uncertainty management becomes crucial for enabling higher levels of delegation. Future tools must help developers maintain project awareness comparable to current team leads.