In the fast-paced world of startups, where every decision can make or break a business, one young entrepreneur's story of AI-driven spending offers a fascinating glimpse into the challenges and opportunities of the modern workplace. Sarthak Dhawan, a 21-year-old co-founder of Turbo AI, shares his experience of accidentally spending $30,000 on AI tokens in a single month, and the simple fix they implemented to avoid a repeat of this costly mistake.
The Costly Mistake
Dhawan's story begins with a month of intense innovation and rapid development. He and his co-founder, Rudy, were in the midst of launching their AI learning tool app, and the pressure to deliver was high. In this context, the $30,000 spent on Claude Code tokens was not a total disaster, but it was a wake-up call. Dhawan explains, "I don't think of that month as a total mistake, though I've learned from it. That was a heavy shipping month, and the spend reflected that. High token months usually mean we're innovating or trying new things."
The Shift in Role
One of the most intriguing aspects of Dhawan's experience is the transformation of his role as an engineer. With the advent of AI code-generation models, his job has evolved from writing code to reviewing it. He notes, "More of my day-to-day is trying to plan and describe how things are working at a high level and reviewing AI-generated code than actually writing my own. It's a lot of vibe-checking stuff as I go." This shift raises a deeper question: as AI takes over more tasks, what does it mean for the skills and knowledge of engineers?
The AI Spending Budget
Dhawan and his team have a unique approach to managing their AI spending. They don't have a set budget for AI token spending, but rather keep a loose eye on it. This flexibility allows them to experiment and innovate without the constraints of a strict budget. However, as the team grows and AI usage increases, costs can creep up. Dhawan mentions, "We have about 10 people on our team, and costs have crept up as more of us use AI for more work. We're fine with it as long as it's driving output. We average around $20,000 a month on AI tooling costs for software development."
The Easy Fix
The $30,000 bill was a result of Dhawan leaving the fast mode setting on Claude, which significantly increased the cost per token. The fix was simple: turning off the fast mode setting. Dhawan reflects, "Switching out of fast mode barely made a difference in the speed of output. Normal mode is plenty fast, and the quality's the same, so it was an easy save that we didn't feel that deeply."
The Token-Saving Strategy
When it comes to saving tokens, Dhawan and his team don't overthink it. They focus on the easy wins, such as defaulting to standard mode, using lighter models for simple tasks, and avoiding dumping whole codebases into context. However, they don't stress over every dollar, believing that if AI makes them more productive, it's worth it for the business in the long run.
The Broader Implications
Dhawan's story raises important questions about the future of work and the role of AI in the workplace. As AI becomes more integrated into software development, what will happen to the skills and knowledge of engineers? Will the codebase become an entity that engineers no longer understand? These are questions that Dhawan and his team are grappling with, and they are not alone. As AI continues to evolve, these issues will only become more pressing.
Conclusion
In the end, Dhawan's story is a reminder that while AI can be a powerful tool for innovation and productivity, it also comes with its own set of challenges. As we continue to embrace AI in the workplace, it's crucial to consider the broader implications and ensure that we are not sacrificing the skills and knowledge of our workforce in the process. From my perspective, Dhawan's story is a call to action for businesses and individuals alike to think critically about the role of AI in the workplace and to ensure that we are using it in a way that benefits us all.
