Walk through the five phases of AI training, choosing an approach at each step to see how real labs differ — and where decisions most move environmental impact. With cited sources throughout.
Walk through data collection, pre-training, post-training, evaluation, and deployment, choosing an approach at each step to see how labs differ in practice. Each choice surfaces key points, discussion questions, and a peer-reviewed or lab-published source. A concluding section highlights where decisions most move environmental impact — grid choice and model size — and what the field still doesn’t disclose.