Amateur Builds Tax Prep Software with AI

Why It Matters

If validated, the prototype proves AI can compress years of software development into weeks, threatening incumbent tax‑software giants and reshaping the low‑code market. It also raises questions about regulatory compliance and consumer trust in AI‑built financial tools.

Key Takeaways

• Journalist built tax app using AI without coding
• App called Telos Tax offers free federal, state filing
• Generates 234,000 lines of code in weeks via Claude
• Operates offline, AES‑256‑GCM encryption, zero data leakage
• Open‑source on GitHub, invites tax professional stress‑testing

Pulse Analysis

The emergence of "vibe coding," where natural‑language prompts drive AI models like Anthropic's Claude to write production‑grade software, marks a turning point for rapid application development. Lizza’s Telos Tax demonstrates that a complex, compliance‑heavy product—traditionally built by hundreds of engineers over multiple years—can be assembled in weeks without any programmer on the team. This democratization lowers entry barriers, enabling entrepreneurs and niche experts to prototype solutions that were once the exclusive domain of Fortune‑500 R&D labs.

Security and privacy are front‑and‑center in Telos Tax’s design. By keeping the entire tax engine in the user's browser and encrypting data with AES‑256‑GCM, the app eliminates cloud‑based data exposure. The optional BYOK (Bring Your Own Key) mode lets power users supply their own Anthropic API key, while an audit log records every interaction with the AI, ensuring transparency about what information leaves the device. Such architecture addresses growing consumer concerns about data harvesting and aligns with emerging regulations that demand minimal data retention for financial applications.

The broader accounting‑software market must now reckon with a potential disruption. If independent auditors confirm Telos Tax’s calculations are accurate, the model could spur a wave of AI‑generated, open‑source financial tools that undercut traditional pricing structures. Established players like Intuit and H&R Block may accelerate their own AI integration or explore partnerships with open‑source communities to retain relevance. However, regulatory scrutiny will intensify, as tax filing software must meet strict IRS standards; rigorous third‑party validation will become a prerequisite for any AI‑crafted solution to gain mainstream adoption.