Once AI becomes part of the production stack, the engineers responsible for deploying it must understand more than code.

Over the weekend, Apple released new research that accuses most advanced generative AI models from the likes of OpenAI, Google and Anthropic of failing to handle tough logical reasoning problems.