Is RAG Dead? Lessons from Building AI for Tax Law with Alex Bowcut - #769
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)Show Notes
As context windows grow into the millions of tokens, many AI practitioners are questioning whether retrieval-augmented generation (RAG) is still necessary. If modern models can ingest entire libraries of documents, why bother with retrieval at all?
In this episode, Alex Bowcut, Head of Engineering at Sphere, explains why the answer depends on the application. Sphere uses AI to automate global tax compliance—an environment where getting the answer right isn’t enough. Every conclusion must be backed by the correct legal citation, and every decision must withstand expert review.
We explore how Sphere built TRAM (Tax Review and Assessment Model), a production AI system that combines retrieval, reasoning models, legal review workflows, reinforcement learning, and deterministic systems to help tax experts move nearly two orders of magnitude faster while maintaining accuracy.
Along the way, we discuss why RAG remains critical in high-stakes domains, how Sphere processes legal and regulatory documents from jurisdictions around the world, retrieval architectures, semantic chunking, dense versus sparse retrieval, expert feedback loops, and the challenges of building AI systems that people can actually trust.
🗒️ Full show notes: https://twimlai.com/go/769.
More from The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
AI agents need the institutional knowledge experienced employees rely on: why a system works a certain way, which policies apply, and what never made it into the documentation. Capturing that knowledge is hard. Keeping it useful as people and agents add more may be harder.
In this episode, Zoe Weil, co-founder and CEO of Sequen, joins us to discuss how ranking systems can use live behavioral signals to optimize for purchases rather than clicks or search relevance alone. Zoe explains why approaches developed at data-rich social platforms don’t necessarily translate to retailers and how Sequen uses scroll speed, hover duration, and other in-session interactions to infer shopper intent without relying on user identity.
In this episode, Diogo Almeida, co-founder and CEO of TypeSafe, joins us to discuss Jev, TypeSafe’s recently released model for bringing fast, reliable intelligence directly into software. We explore the idea of “machine-native intelligence” and why Diogo believes models optimized for generating text are poorly suited to many of the decisions required for real-world automation. He explains how Jev differs from traditional classifiers and LLM-based approaches, the role of reinforcement learning from calibrated decisions (RLCD), and why calibration and reliability are central to making AI useful as a software primitive. We also discuss the relationship between models and code, why Diogo believes AI systems should become more engineered rather than relying on a single model to do everything, and how Jev-like models could reshape agents, tool use, and the architecture of AI-powered software.
AI systems have gone from struggling with grade-school math to helping solve research problems that have resisted mathematicians for decades, including Navier-Stokes.
Voice AI has gotten remarkably good, but natural conversation remains a high bar. Small delays, awkward interruptions, or the wrong tone can quickly break the illusion—and adding vision and visual presence only raises the stakes.