Stop using ChatGPT for everything: I use these AI models for research, coding, and more (and which I avoid) (www.zdnet.com)

In a recent exploration of various AI models for specific tasks, the author emphasizes that not all AI models are created equal, and understanding their unique strengths can greatly enhance productivity. With the rapid advancements in generative AI, it’s essential for users to focus more on their workflow rather than obsessing over model versions. The distinctions among models like GPT-5.1, Opus 4.5, and Gemini 3 become clearer as the author shares practical applications. For instance, Google’s NotebookLM excels in creating audio explainers from dense documents, while advanced coding tasks are efficiently handled by OpenAI's Codex and Claude Opus 4.5. The article also highlights the increasing prevalence of AI subscription costs across various software applications, urging users to be strategic about which models and tools they invest in based on their needs. The author suggests that the AI models should be chosen based on specific tasks—such as using Karakeep for keyword generation without additional subscription fees or Notion AI for article summarizations—rather than defaulting to popular options like ChatGPT for every scenario. This nuanced approach not only saves time but can also significantly increase the quality of outputs in research, coding, and content management.