Insights & editorial

Thinking in systems and code with aesthetics.

Technology3 min

A large compute facility can expand AI supply. GPU counts alone do not determine the experience of using a coding agent.

Technology3 min

Access to weights does not remove infrastructure, maintenance or licensing requirements. Compare a managed service with operating the model yourself.

Technology3 min

Evaluate the work your team needs to finish, with recorded configurations and consistent criteria for quality, cost and review effort.

Technology3 min

A large context window does not replace a project map, acceptance criteria or a durable record of decisions made during a task.

Technology3 min

Organize the objective, context, permissions and verification when using a terminal agent. The result still needs to withstand code review.

Business2 min

Available weights create deployment options, but the decision also involves licensing, capacity, observability and operational responsibility.

Technology2 min

Compare Claude models in your own repository by recording quality, interventions, time and cost per accepted task.

Technology2 min

Define context, boundaries and validation for a real Claude Code task, from initial investigation to the diff and tests.

Technology2 min

A larger window can bring related documents together. Testing the task shows whether the extra context improves the answer.

Technology2 min

Use independent tasks, verifiable deliverables and accountable integration to test whether multiple agents improve the outcome.

Technology3 min

Keep the names Sol, Terra and Luna unchanged. Test the GPT-5.6 family without confusing the model, reasoning effort, working tool and cost.

Technology3 min

Adoption includes the environment, access, review and delivery criteria. Choosing a model is only one part of the work.