Osaurus makes the case for running AI locally on the Mac

Podcast puts local AI on the Mac in focus

The latest episode of the Apple @ Work podcast, published by 9to5 Mac on 10 November 2026, features Terence Pae, the co-founder and chief executive of Osaurus. The conversation centres on a simple idea: companies may be better served by running AI locally on the Mac rather than sending data to cloud services.

Osaurus describes itself as a free, first-party AI harness for the Mac. In plain terms, a harness of this kind is software that lets a computer run AI models on its own hardware, without a constant connection to outside servers.

Why local AI matters to organisations

The pitch for local AI rests on control. When AI tasks run on a device, information does not have to leave the building, which can matter to organisations with strict data policies. Apple silicon Macs are also built around efficient processors, which the podcast discussion suggests makes them a practical home for on-device AI work.

According to 9to5 Mac, the episode explores these arguments in the context of workplace deployments, where IT teams weigh convenience against privacy and cost. The details of those arguments sit in the audio discussion itself, and this article reports only what the published summary confirms.

What readers should keep in mind

Osaurus being free and first-party is a notable detail, because it suggests the tool is meant to sit close to the Mac platform rather than act as a third-party add-on. Beyond that, the published material does not include benchmarks, supported models or enterprise pricing, so readers should treat the episode as a discussion of direction rather than a product review.

Businesses curious about the approach can listen to the episode on 9to5 Mac, and anyone evaluating the software should check the vendor's own documentation before deployment.