The Latent Space podcast features Jacob Lauritzen, chief technology officer of legal AI company Legora, in an episode on building AI products for lawyers. According to the episode notes, the conversation covers three areas: evaluating AI systems without relying on full execution traces, running reviews across tabular datasets of up to a million cells, and whether the billable hour can remain the core business model for law firms as AI handles more document work. The show notes do not detail Lauritzen's specific positions or product figures, so the full discussion should be checked against the episode itself. The episode sits at the intersection of two pressures facing legal services. Firms and their software vendors are trying to build AI tools that are accurate enough to rely on in client work, which makes evaluation methods central to product development. At the same time, if AI reduces the hours needed for tasks such as document review, the hourly billing model that underpins many firms comes under strain, which is why the topic has drawn attention across the sector. For Australian businesses and professionals, the episode itself changes nothing immediately; it is a technical and strategic discussion rather than a product launch or regulatory change. Law firms and in-house teams considering legal AI tools may find it useful to ask vendors how they test accuracy, where client data is stored and processed, and whether the tool meets their obligations under the Privacy Act 1988 and state professional conduct rules. Fee arrangements may also need review if clients begin questioning hourly charges for AI-assisted work.