Until 18 months ago, I wrote every line of code myself. Today, AI writes 80% of my initial implementations while I focus on architecture, review, and steering multiple development threads simultaneously. This isn't another "AI will change everything" post. This is about the messy reality of integrating AI into production development workflows: what actually works, what wastes your time, and why treating AI like a "junior developer who doesn't learn" became my mental model for success. The backstory: We run monthly engineering workshops at Sanity where someone presents what they've been experimenting with. Last time was my turn, and I showed how I'd been using Claude Code. This blog post is from my presentation at our internal workshop (10-min recording below). My approach to solving code problems has pivoted four times in my career: For the first 5 years, I was reading books and SDK documentation. Then 12 years of googling for crowd-sourced answers. It was 18 months of using Cursor for AI-assisted coding And recently, 6 weeks of using Claude Code for full AI delegation Each transition happened faster than the last. The shift to Claude Code? That took just hours of use for me to become productive. Here's what my workflow looks like now, stripped of the hype. I use AI mostly "to think with" as I'm working with it towards the code that ends up in production. Forget the promise of one-shot perfect code generation. Your job as an engineer is to find the best solution for the problem, not just write a bunch of code. Claude builds context about your system You identify the actual challenges The code is usually completely wrong Then you take the learnings from this attempt and feed it back. Claude understands the nuances You've defined concrete approaches Half the time, it's still unusable Claude implements something we can iterate on and refine You constantly review and course-correct This becomes your starting point, not your final code This isn't failure; it's the process! Expecting perfection on attempt one is like expecting a junior developer to nail a complex feature without context. The biggest challenge? AI can't retain learning between sessions (unless you spend the time manually giving it the "memories"). So typically, every conversation starts fresh. My solutions: Create a project-specific context file with: Architecture decisions Common patterns in your codebase Gotchas and workarounds Links to relevant documentation Thanks to MCP integrations, I can now connect my AI to: Linear for ticket context Notion or Canvas for documentation Non-production databases (only with read access!) for data and data structures Your actual codebase (obviously) Github (get useful background context from older PRs) Without this context, you're explaining the same constraints repeatedly. With it, you start at attempt two instead of attempt one. How my Claude Code is connected to other tools mainly for gaining context I run multiple Claude instances in parallel now, it's like managing a small team of developers who reset their memory each morning. Key strategies: Never parallelize the same problem space (it's easy to lose track and confuse the different problems you're solving) (it's easy to lose track and confuse the different problems you're solving) Track everything in Linear (or whatever project management tool you use) (or whatever project management tool you use) Explicitly mark human-edited code (AI gets confused about what it wrote versus what you modified) Writing code is one part of the job, but so is reviewing code. Adopting AI has evolved my code review process as well. Catches missing test coverage Finds obvious bugs Suggests improvements This saves me and my peers time and extra rounds. At Sanity, our policy is that the engineer is responsible for the code they ship, even if it's AI generated. I want to make sure that I ship: A maintainable codebase Sound architecture decisions Business logic correctness Good integration points They rarely know which code is AI-generated Quality bar remains the same The key take away: I'm more critical of "my code" now because I didn't type out a lot of it. No emotional attachment means better reviews. We're testing Slack-triggered agents using Cursor for simple tasks: 2 successes with business logic fixes 1 failure with CSS layouts Current limitations: No private NPM package access It passes unsigned commits Bypasses normal tracking The Cursor agent works best for simple tasks But the potential? Imagine agents handling your backlog's small tickets while you sleep. We're actively exploring this at Sanity, sharing learnings across teams as we figure out what works. Let's talk money. My Claude Code usage costs my company not an insignificant percent of what they pay me monthly. But for that investment: I ship features 2-3x faster I can manage multiple development threads I spend zero time on boilerplate and repetitive code The ROI is obvious, but budget for $1000-1500/month for a senior engineer going all-in on AI development. It's also reasonable to expect engineers to get more efficient with AI spend as they get good with it, but give them time. Not everything in AI-assisted development is smooth. Here are the persistent challenges I find myself in: The learning problem AI doesn't learn from mistakes. You fix the same misunderstandings repeatedly. Your solution: better documentation and more explicit instructions. The confidence problem AI confidently writes broken code claiming that it's great. Always verify, especially for: Complex state management Performance-critical sections Security-sensitive code The context limit problem Large codebases overwhelm AI context windows. Break problems into smaller chunks and provide focused context. The hardest part? Letting go of code ownership. But now I don't care about "my code" anymore; it's just output to review and refine. This detachment is actually quite liberating! Faster deletion of bad solutions More objective code reviews Zero ego in refactoring If a better AI tool appears tomorrow, I'll switch immediately. The code isn't precious; the problems we solve are. If I were to give advice from an engineer's perspective, if you're a technical leader considering AI adoption: Let your engineers adopt and test different AI solutions: AI-assisted coding is a skill that you have to practice to learn. Start with your most repetitive tasks: that's where AI shines immediately. Budget for experimentation: the first month will be messy. Adjust your review processes: AI code needs different scrutiny. Document everything: Great context is your efficiency multiplier. The engineers who adapt to the new AI workflows will find themselves with a new sharp knife in their toolbox: They're becoming orchestrators, handling multiple AI agents while focusing on architecture, review, and complex problem-solving. Pick one small, well-defined feature. Give AI three attempts at implementing it. Review the output like you're mentoring a junior developer. That's it. No huge transformation needed, no process overhaul required. Just one feature, three attempts, and a honest review. The future isn't about AI replacing developers. It's about developers working faster, creating better solutions, and leveraging the best tools available.