ChatGPT deserves a more generous comparison than the original brief gave it. Projects are available across free and paid plans, can keep chats, files, and instructions together, and use memory to carry relevant context forward. Scheduled tasks can provide reminders and recurring follow-up. In the current U.S. rollout, ChatGPT Health can also read approved Apple Health and medical-record data and use it to ground health conversations. The claim that every new chat simply starts from zero is no longer accurate.
Its breadth is the advantage. ChatGPT can discuss a culturally specific meal, explain a health concept, compare ingredients, work around a budget, search for current information, and reshape an initial plan through natural conversation. OpenAI itself teaches meal planning as a practical use case. A person who enjoys directing the process can build a genuinely useful nutrition workflow around it, often without paying anything.
The remaining distinction is product structure, not intelligence. ChatGPT does not document a built-in weekly food system that keeps recipes, individual household portions, a live grocery list, plan-versus-actual eating, and check-in-driven changes in one nutrition workflow. Research on AI meal plans is also mixed: one 2025 study found ChatGPT was the most precise of three chatbots at meeting calorie targets, while a 2026 study under ordinary-user conditions found all tested language models produced plans 284 to 546 calories below target. Eat Like This narrows the problem, constrains the recipes, and turns the output into an ongoing product state.