I Built Two Apps with Just My Voice and a Mouse - Are IDEs Already Obsolete?

Why It Matters

If AI can handle coding, testing, and bug fixing via simple prompts, the demand for complex IDEs may decline, reshaping the software‑tooling market. This shift could lower development costs and accelerate product cycles.

Key Takeaways

• AI-driven "vibe coding" replaces edit/debug steps
• Terminal + voice dictation handles full app development
• IDEs limited to build and deployment actions only
• Multi‑platform Apple apps built in hours, not months
• Productivity gains could reshape developer tooling market

Pulse Analysis

The rapid maturation of large‑language‑model assistants is redefining how developers write code. Tools like Claude Code, GitHub Copilot and Google Gemini can interpret natural‑language instructions, generate syntactically correct code, and even suggest fixes when prompted. This capability transforms the classic edit‑build‑test‑debug cycle into a streamlined "instruct‑build‑test‑guide" loop, where the developer’s role shifts from manual typing to strategic direction. By moving the interaction surface to a chat or terminal window, developers can stay in a single context, reducing cognitive load and eliminating the need for heavyweight feature sets such as syntax highlighting or breakpoint management.

Gewirtz’s two‑hour experiment illustrates the practical upside of this paradigm. Using a mouse‑mapped return key and Wispr Flow for voice dictation, he directed an AI to create and iterate on two distinct iOS/macOS/watchOS applications entirely from iTerm2. The AI handled boilerplate code, NFC integration, and on‑device image analysis, while the developer supplied high‑level prompts and corrective feedback. The only IDE involvement was a single command to trigger a build and push the binary to TestFlight. This workflow cuts down on context switching, accelerates prototyping, and makes development feasible with one hand free—an advantage for accessibility and remote work scenarios.

The broader industry impact could be profound. IDE vendors like JetBrains, Microsoft and Apple may need to pivot toward tighter AI integration, offering lightweight launchers or AI‑centric extensions rather than full‑stack environments. Enterprises might adopt terminal‑first, AI‑augmented pipelines to reduce licensing costs and streamline onboarding. However, challenges remain: AI‑generated code still requires rigorous security review, and complex debugging of performance‑critical systems may demand traditional tools. As AI assistants improve, the balance between convenience and control will shape the next generation of developer platforms.