AI Won't Take My Job. I Quit First.

AI is rapidly redefining the software developer's core responsibilities. Once primarily code-writers, developers now evolve into system-architects and problem-definers, leveraging AI for foundational tasks like boilerplate generation. Elliot, a veteran developer since 2007 and host of Dreams of Code, exemplifies this shift, utilizing AI for generating code stubs, defining public interfaces, and aggressive Git staging. He even switched from Claude Code to Codex, highlighting specific tool preferences in this new landscape.

Recognizing this paradigm shift, Elliot proactively left his traditional development role. He transitioned to a full-time content creator for his popular YouTube channels, Dreams of Code and Dreams of Autonomy, accumulating hundreds of thousands of subscribers. This strategic move allows him to build skills in areas AI cannot easily replicate, emphasizing human expertise and creative problem-solving over rote coding.

Traditional learning methods, particularly paid coding courses, are also experiencing a significant decline. Developers increasingly bypass lengthy tutorials, instead turning directly to AI tools like Claude Code and Codex for instant knowledge and rapid code generation. This instant access reshapes how new skills are acquired, signaling the "death of coding tutorials" as extensively discussed on the Better Stack Podcast Ep.. 16. The emphasis shifts from memorizing syntax to effective AI prompting and integration.

Elliot, the insightful developer behind the Dreams of Code and Dreams of Autonomy YouTube channels, articulated a bold strategic bet on the Better Stack Podcast Ep..: he’s going "all-in on Rust for 2026." This isn't a casual preference but a calculated move to future-proof his skills against the backdrop of rapidly evolving AI capabilities.

Rust’s core strengths in memory safety and exceptional performance position it as a critical language for complex, high-stakes applications. These attributes are indispensable for projects like Elliot’s ongoing cross-platform video editor, which demands robust, low-latency execution and careful management of intricate components, including extensive use of unsafe FFI.

Industry adoption of Rust mirrors this strategic foresight. The language is increasingly foundational for systems programming and high-performance backend services where reliability is paramount. As of May 2025, 45% of organizations utilize Rust for non-trivial production systems, marking a significant 7-percentage-point increase from the prior year and affirming its status as the most-admired programming language with an 83% score.

Current large language models (LLMs), while adept at generating boilerplate or public interfaces, consistently struggle with the nuanced, error-prone domains Rust excels in. The guarantees of memory safety and concurrency control inherent to Rust’s design remain beyond the current generative capacity of AI, ensuring a human-centric demand for expert developers crafting high-reliability code.

Elliot, a seasoned developer and content creator, leverages AI not for wholesale code generation but as a precision tool within his daily workflow. He integrates models for generating code stubs, defining public interfaces, and enabling aggressive Git staging. This targeted approach streamlines development, allowing him to offload repetitive structural tasks and focus on higher-level problem-solving and system architecture.

Initially relying on Claude Code, Elliot ultimately transitioned to Codex, citing superior workflow integration and task-specific performance. His preference for terminal-based AI interaction, rather than feature-laden IDE extensions like Cursor, underscored the shift. He noted that such tools, while powerful, often "throw all of its everything it wants to sell me in my face," hindering simple code review and editing. For deeper insights into its capabilities, explore the OpenAI Codex documentation.

Elliot articulates a controversial perspective: large language models are reaching diminishing returns for core coding tasks. He views them evolving into indispensable utility tools, rather than revolutionary replacements for human developers. This framing positions AI as an augmentative force, optimizing specific development phases without fundamentally altering the need for human architecting and problem-solving. This shift signifies AI's role as a powerful assistant, not an autonomous creator, demanding a re-evaluation of its long-term impact on developer roles.

Developers are increasingly returning to minimalist tools, embracing a terminal renaissance. Elliot, known from Dreams of Code, boasts a long history with Neovim. He first adopted Vim at university, crafting a C++ compiler over SSH into a Unix box, a skill cemented in his first finance job. This early immersion established Vim as his primary editor, capable of "pretty much anything."

Modern IDEs, however, often overwhelm with intrusive AI features. As James from the Better Stack Podcast Ep.. notes, tools like Cursor feel "filled with other AI stuff," constantly "popping up in my face." This relentless auto-completion and suggestion engine disrupts the flow, turning powerful environments into cluttered distractions.

Elliot finds liberation in this pared-down approach. He leverages AI, specifically Codex via the terminal, for foundational tasks like generating stubs and defining public interfaces. This allows AI to handle initial scaffolding, while he retains precise control for refinement within the uncluttered environment of Neovim.

This division of labor optimizes the development process. Developers can use CLI tools for aggressive Git staging and rapid code generation, then switch seamlessly to a keyboard-driven editor for meticulous edits. The terminal becomes a clean workspace, free from the constant noise of AI attempting to anticipate every keystroke, ensuring the human developer remains firmly in command.

Why did developer Elliot quit his job because of AI?

He quit proactively to adapt to the AI revolution, choosing to focus on future-proof skills like Rust and content creation rather than wait for AI to potentially displace his traditional role.

Why is Rust considered a good language to learn in the age of AI?

Rust offers memory safety without a garbage collector and high performance, making it ideal for complex systems-level programming where AI-generated code often lacks the necessary nuance and reliability.

What's the key difference between Claude Code and Codex in a developer's workflow?

Based on Elliot's experience, he switched to Codex for its utility in generating stubs and public interfaces. This suggests Codex integrates better into a rapid, terminal-based workflow, while Claude is often noted for more complex reasoning and architectural planning.

Are terminal editors like Neovim still relevant with AI?

Yes, many developers prefer Neovim for a distraction-free environment. They use AI tools to generate code but value the control and focus of a minimalist editor for reading, navigating, and refining it.