How Cursor helped me fall back in love with coding
I took a five year break after leaving Google. Here's why AI-powered code editors have made it impossible for me to not come back to my software engineering roots.
In August 2024, my social feeds were on fire with excitement about AI coding tools—Cursor, GitHub Copilot, Replit Assistant.
Video after video, tweet after tweet, people were calling them game-changers.
Curious but skeptical, I texted a friend at Airbnb: “Do these AI tools really live up to the hype?” His reply—“oh absolutely”—was the final nudge I needed.
So what would I build first?
Well, I had been wanting a simple app to nudge me in doing my daily affirmations.
At the time, I was getting pretty into…
Identifying a single focus of the day like “I will deliver on my role at <insert consulting client> and make the people I work with proud”; and
Writing this focus in a journal, three times in the morning, six times in the afternoon, and nine times in the evening.
But I wanted to use my voice to speak these affirmations. And I wanted a way to easily motivate my self in an almost game-like format to complete all three sessions for a day, and keep the momentum going day after day.
So I did a bit of research on existing tools, and when nothing matched what I wanted, I thought: Why not build it myself?
That same evening, I fired up Cursor, an AI-powered code editor, for the first time. I hadn’t coded seriously in years, so I expected it to be a slog. But from the first hour, it felt like entering a new world.
Every time I hit a roadblock—“How do I structure this database? How should I handle voice input?”—Cursor was right there with suggestions, boilerplate, and quick refactors.
By the end of that evening, I had a functional prototype.
I was floored.
The next day, I gave my new app a spin. It listened to my voice, tracked if I’d completed my morning, afternoon, and evening sessions, and gave me a streak point for completing the day successfully.
But the real magic wasn’t just shipping a first version—it was using and improving it over time. I didn’t just build a shell of an app and leave it on a digital shelf; I opened that app up for three sessions a day for 40 days, speaking my affirmations into it, refining it as I went. A friend of mine hopped on too, and they also kept a 40-day streak, becoming my unofficial alpha tester and pushing me to add clever features.
I was hooked. Not just on using the app, but on building again.
In the weeks that followed, I spun up other projects, including one that would eventually go live on the Apple App Store, Google Play Store, as well as on the web with full functionality.
This wasn’t like my days at Google Cloud from 2015 to 2019. Back then, shipping code with Google’s internal devtools was a decent experience—there were countless code samples to learn from and brilliant engineers a desk away—but it often felt like navigating a dense forest. Each step forward required careful deliberation, and all too often, we’d hit demoralizing dead ends.
Now, with AI-powered coding tools, not to mention AI-supercharged search engines like Perplexity, everything felt more fluid. When I needed to center a div, it happened in seconds. When I had a weird syntax error, it almost always got fixed after one or two tries.
The software engineer’s flow state. There’s nothing quite like it. And AI is making these states way more common.
That warm August evening was a turning point. It wasn’t just about building a single app; it was about discovering a new way of creating, iterating, and solving problems.
I’m back in my software engineering flow. But this time is totally different and better.
The Era of AI-Supercharged Coding Is Here
“[AI coding tools] bring joy back to software development… [They help developers] stay in flow.”
— Microsoft CEO Satya Nadella
Generative AI tools have fundamentally reshaped the coding process, allowing developers to focus on problem-solving, user experience design, and high-level systems architecture instead of repetitive tasks.
Rapid prototyping, quick iteration, and seamless refinement are now the norm, accelerating development like never before.
It may take a few weeks for a pre-AI-first developer to adjust, like it did for me. But the crossover point—where AI makes you more productive than working without it—arrives sooner than you’d expect if you commit to using it every day.
This transformation isn't theoretical—it's already happening. As Sundar Pichai, CEO of Google, highlights:
“Today, more than 25% of all new code at Google is created by AI, which is then reviewed and approved by engineers. This enables our engineers to accomplish more tasks and accelerate their work.”
But it’s not just the tech giants. A 2024 GitHub survey revealed that nearly all developers surveyed are using AI coding tools, with most reporting major boosts in productivity and code quality.
Take PDF.ai—a one-person company that scaled to $1.5 million in annual recurring revenue (ARR) in just over a year.
Or Midday, a small business automation platform powering 9,000 freelancers and small enterprises with just two team members.
AI-powered code editors let the founder build, iterate, and scale with remarkable speed.
This kind of solo or nano-sized team-driven success will only become more common as AI continues to lower the barriers to building and scaling software products.
And it’s not only about speed—it’s about empowerment. Solo founders, small teams, and independent developers can now use tools that were once reserved for industry titans.
After shipping several AI-assisted projects myself, I realized the techniques and practices I’d tested were only the tip of the iceberg.
That’s why I’m joined and am currently participating in GauntletAI, an intensive engineering bootcamp led by Austen Allred of BloomTech, to fully explore what these tools can enable. It’s the first Gauntlet cohort, and the educational team is figuring out everything as AI evolves in real time. But it’s been an inspiring experience so far.
Why This Matters
What excites me most is how the “surface area” of solvable problems has exploded.
AI-powered development has dramatically cut both the time and cost of building software, making it viable to tackle small-scale or specialized challenges that used to seem too minor or unprofitable.
Communities can now have digital tools made just for them, and individuals can create highly personalized apps—things that would’ve seemed out of reach and impractical only a few years ago.
Even more remarkable is that technology once locked away in the world’s top tech labs is now available to anyone with a laptop and an internet connection.
Take regenerative farming, for example. Imagine a small or medium scale farmer using AI tools to map soil quality, track weather, and optimize planting for better yields. Soon, they’ll be able to build custom systems tailored to their land.
And it wouldn’t have to be a solution created by a venture-backed startup. They could could affordably get someone to build something precise for their own land, experiment season by season, and continually refine their approach.
The real magic kicks in with open-source. Farmers could share their data—soil metrics, crop yields, weather patterns, and regenerative methods—with shared AI models.
In return, these farmers wouldn’t just get smarter tools informed by aggregated collective insights; they could also share in the value those models generate.
But agriculture is only part of the puzzle. This same interplay—local and niche creation, collective intelligence, and shared value—applies to every corner of society.
From teachers crafting AI-driven lesson plans to doctors devising personalized treatment options, we’re poised to unlock abundance and creativity on a scale we can barely imagine.
The democratization of building real-world tools isn’t some far-off vision; it’s unfolding right now.
And as it does, the ripple effects across every sector of society will be transformative.
My Reincarnation as a Software Engineer
I enjoyed coding at Google, but at some point, I shifted focus to projects that let me tap into other strengths and passions—like building communities, producing events, writing and content creation, and crafting creative marketing campaigns.
But let me reiterate: I’m back in my software engineering flow.
Since August, I’ve built (or contributed to) several tools that I not only use regularly but that have also found traction with at least a few early adopters who find value in them:
BlessOut – A voice-centric e-card app for soulful digital communication, currently in alpha on web, Android, and iOS.
Jakki – A Telegram bot that helps a community co-run an X or Farcaster account by upvoting posts and running caption competitions.
MyFinancialFuture – A planning tool for sketching out high-level financial goals in an increasingly expensive world.
Flow State - A crypto-native streaming funding rewards platform and co-op. I helped launch the project and will contribute to the codebase when I can.
A voice-first, multi-platform life journal with an AI agent sidekick who supports you as you embark on your hero’s journey. More on this to come!
These aren't just experiments—they're tools I have come to rely on. I use BlessOut multiple times a week to keep my most important relationships warm, my life journal about a dozen times a day to document and reflect on daily learnings and experiences, Jakki is already powering social accounts for several token-centric communities, my MyFinancialFuture printout posted above my desk motivates me every day to work hard and earn my keep, and Flow State is just humming in the background powering multiple streaming funding rounds for crypto ecosystems.
What’s wildly exciting to me is how quickly the first versions of each of these tools came together. Just a few years ago, the labor usually required to create an initial proof of concept was weeks of painstaking effort, but now it may only take days—if not just hours—using AI-powered development tools like Cursor.
Ideas that used to live only in my head are suddenly only a hop and skip away from being real. And this increase in agency isn’t just available for me—it’s available for anyone willing to dive into the world of AI-first development.
I can’t stress this enough:
The surface area of what’s worth building—and what’s actually buildable—has expanded in ways we’re only beginning to grasp.
If you’d like to try BlessOut, DM me for an invite code. Preview here below:
If You’ve Ever Thought About Coding, Now Is the Time
Whether you’re new to coding or a seasoned developer, there’s never been a better time to dive into AI-supercharged dev tools. The barriers to building something cool and useful are lower than ever, and where things are going, the potential will soon be limitless.
I’ll be sharing my wins, mistakes, and discoveries related to AI-first development on X and here on Substack—so subscribe and follow if you aren’t already. If you build something cool with these tools, let me know so we can learn from each other.
For now, here are a handful of AI-first software engineering creators who I encourage you to subscribe to:
Riley Brown - “the most technical non-technical person on the internet” who uploads end to end tutorials of how to build apps with Replit and Cursor
Michael Shimeles (Ras Mic) - a talented freelance developer who has won major hackathons and also uploads to end to end tutorials of building apps with AI coding tools
Zaid Mukaddam - a self taught developer who recently had the top trending repo on GitHub, with his Scira open source search engine project.
Greg Isenberg - teaches people how to build SaaS companies with little to no budget, uploading 1-2 hour tutorials, recently focused on AI agents
Theo.gg - a popular YouTuber engineer who shares insights on building AI applications and full-stack TypeScript projects
Sawyer Hood - super talented dev creating a lightweight version of Claude Computer Use via his DoBrowser plugin
serafim - a designer turned “design engineer” who shipped an awesome component library platform that went #1 on Product Hunt this past weekend.
Shaw - creator of elizaOS, a massively popular social AI agent framework, and great educator about the potential of AI-first development. Check out his AI Agent Dev School. He made it good for coding novices.
Nader Dabit - amazing developer relations leader who has been hosting AI agent demo days and creating practical AI x crypto tutorials
If you’ve been on the fence about coding—or stepped away for a while—there’s never been a better moment to jump in.
Welcome to the Golden Age of software engineering—what will you build now that it’s begun?
🔥
Thank you for sharing. I love hearing about what you’re up to and what you’re building. I’m curious about the journal - keen to test it out.