Patrick Collison argues that modern development tools are a step back from older integrated environments. He envisions a future where hovering over code reveals live profiling, logging, and even the most common production values for variables, thus tightly merging the editor with the runtime environment.
Related Insights
Prototyping directly in the production environment makes high-quality interactions achievable without extensive resources. This dissolves the traditional design dilemma of sacrificing quality for speed, allowing teams to build better products faster.
New IDEs like Gastown, with roles like 'overseer' and 'mayor' managing AI agent 'convoys,' reveal the developer's future. The job is becoming less about writing code line-by-line and more about high-level orchestration, prompting, and reviewing the output of specialized AI agents to complete complex tasks.
Despite the rise of terminal-based AI, IDEs remain essential because source code is meant for human consumption. Visual interfaces are the best way for developers to review, understand, and build context around what AI agents produce, preventing the 'death of the IDE'.
The lines between IDEs and terminals are blurring as both adopt features from the other. The future developer workbench will be a hybrid prioritizing a natural language prompting interface, relegating direct code editing to a secondary, fallback role.
The current model of separate design files and codebases is inefficient. Future tools will enable designers to directly manipulate production code through a visual canvas, eliminating the handoff process and creating a single, shared source of truth for the entire team.
Instead of becoming obsolete, IDEs like IntelliJ will be repurposed as highly efficient, background services for AI agents. Their fast indexing and incremental rebuild capabilities will be leveraged by AIs, while the human engineer works through a separate agent-native interface.
The user interface and features of the coding environment (the 'harness'), like Cursor or the Codex desktop app, significantly impact AI model performance. A poor experience may stem from an immature application wrapper rather than a flaw in the underlying language model, shifting focus from model-vs-model to the entire toolchain.
The next IDE evolution will transform the codebase into a dynamic 'metadata backbone'. By capturing a continuous history of edits and conversations, it will allow all context—discussions, decisions, feedback—to be permanently anchored to specific lines of code, unlike today's static, snapshot-based Git workflows.
Patrick Collison finds it surprising that programming paradigms haven't fundamentally changed in decades, despite an explosion in the number of developers. He notes that core ideas like integrated development environments originate from the 70s and 80s, suggesting the 'aperture of experimentation' has been disappointingly narrow.
For over a decade, software development fragmented into siloed roles (PM, Design, Eng) with their own tools. AI code editors are collapsing these boundaries by creating a unified workspace where a single "maker" or a streamlined team can build, iterate, and ship, much like in the early days of computing.