Stop using ChatGPT for everything: My go-to AI models for research, coding, and more (and which I avoid)

A ZDNET article provides practical guidance on choosing AI models and apps based on task and workflow rather than version numbers, comparing tools like ChatGPT (GPT-5.x), Gemini 3, Claude Opus 4.5, Midjourney, NotebookLM, Notion AI, Paraspeech, and agentic coding tools (Codex, Claude Code). It highlights cost dynamics (stacked subscriptions vs. API usage), strengths and weaknesses across use cases (diagrams/images, coding, research, speech recognition, knowledge management), and notes mixed tool performance depending on chatbot vs. agentic contexts. The piece also shares tools the author avoids (Perplexity, Copilot, Grok), touches on “Deep Research” tiers, and emphasizes selecting applications that fit specific workflows to maximize value.