There are a lot of AI coding applications out there, and as impressive as large language models and the agents they enable have become, many of the most recent developments in AI‑assisted development have been in the software that manages those models, not just the models themselves.

Claude Code keeps a lean harness, a layer that sits between a model and a developer’s project, deciding what the model sees and how it acts, but it trusts the model to evolve rather than building opinionated tools.

Augment Code, led by Vinay Perneti, takes the semantic route: it pre‑indexes a repository with embeddings, a retrieval model and a vector database, then pulls conceptually relevant code in sub‑milliseconds. That gives it a 33 % token‑efficiency boost over Claude Code on private codebases.

Perneti argues that even as models grow smarter, they still lack context, and that a well‑built harness can decide where to spend a token budget. Thus, intelligence and context together are the key to higher‑quality outcomes.