Shark Tank judge Anupam Mittal built a fully functional AI-powered content system over a single weekend without writing a line of code, then switched from Anthropic's Claude to Elon Musk's Grok after discovering it cost one-tenth as much while delivering comparable results. The experience highlights how rapidly artificial intelligence is becoming accessible to non-technical professionals and how cost differences between AI models are reshaping real-world adoption decisions.

Mittal, founder of Shaadi.com, documented his experiment on X (formerly Twitter), revealing that he created what he calls eight AI "employees" with names like Manish, Chandan, and Lyndon. The system automates content generation, a task that traditionally required either hiring human workers or learning to code. "Built an automated content system in few hours this wknd. Craazzzyy coz I never coded," Mittal wrote, emphasizing his lack of programming background.

Why Did Mittal Switch From Claude to Grok?

When asked which AI models powered his system, Mittal revealed he initially started with Claude, Anthropic's flagship large language model (LLM), which is widely regarded as one of the most capable AI assistants available. However, after comparing the two tools side by side, he made the switch to Grok, xAI's AI chatbot developed under Elon Musk's company. The deciding factor was economics: Mittal found Grok to be significantly faster and cost approximately one-tenth the price of Claude.

"Started with Claude but finding Grok faster and 1/10th the cost," Mittal stated.

Anupam Mittal, Shark Tank Judge and Founder of Shaadi.com

This cost differential is substantial in the context of AI operations. For businesses running continuous AI workloads, a 90% reduction in per-token pricing (the cost to process words through an AI model) can translate to significant savings at scale. Mittal's experience suggests that xAI's pricing strategy is becoming competitive enough to pull users away from established players, even when those alternatives have strong reputations for quality.

How to Build an AI Agent System Without Coding Experience

• Start with accessible AI tools: Use no-code or low-code AI platforms that allow you to connect existing AI models without writing traditional software code, making automation accessible to non-programmers.
• Compare models by cost and speed: Test multiple AI models on your specific use case to evaluate both performance and pricing, as different models excel at different tasks and price points.
• Validate output quality before scaling: Run small tests to ensure the AI system produces acceptable results for your needs before deploying it to handle critical business functions.

When asked about the quality of his automated system, Mittal reported that it was performing exceptionally well. "Amazing... working like a charm... so far. And it's multiplayer," he noted, suggesting the system can handle multiple concurrent tasks or users. This feedback is significant because it indicates that cost savings did not come at the expense of functionality.

Mittal

What Does This Reveal About AI Adoption in 2026?

Mittal's experiment reflects a broader shift in how professionals are approaching artificial intelligence. Rather than viewing AI as a specialized tool requiring deep technical expertise, he demonstrated that modern AI platforms have become accessible enough for business leaders to experiment with and deploy independently. His reflection on the pace of change captures the sentiment many are experiencing: "Different world in a year. I feel it in my bones now," he wrote.

His

The speed at which Mittal built his system,"in few hours this wknd",underscores how AI development tools have matured. What once required months of planning, hiring, and coding can now happen in an afternoon. This democratization of AI capability is reshaping expectations about what business leaders should be able to accomplish with technology.

Mittal's cost comparison also signals that the AI market is becoming more price-competitive. While Claude remains a capable model, the emergence of alternatives like Grok that offer comparable performance at a fraction of the cost is forcing established players to reconsider their pricing strategies. For businesses evaluating AI tools, this competition creates an opportunity to reduce operational costs without sacrificing quality.

The practical implications are clear: non-technical professionals now have the tools and economic incentives to build AI-powered systems independently, and cost efficiency is becoming as important as raw capability when choosing between competing AI models. As Mittal's experience demonstrates, the barrier to entry for AI automation has dropped dramatically, and the financial barrier is dropping just as fast.