Give the thinking model a task to plan. Return JSON {"goal":"goal and limitations","plan_request":"what exactly are you asking to plan (stages, order, deadlines)","dependencies_request":"how are you asking to take into account dependencies and the critical path (chain of tasks that determines the overall deadline)","risks_request":"how are you asking to identify risks and points of failure"}
Planning is a strong point of reasoning, but only if you set boundaries. Give purpose, resources and limitations; ask for stages with order and deadlines, dependencies (“what can’t be started until something else is ready”), and the critical path—the chain that defines the overall deadline. First trick: ask not only for a plan, but also for a rationale for the order - why this first and then that. This shows the train of thought, and it can be challenged: suddenly the site does not have to be ready before the lessons are recorded. The second trick is not obvious: ask “what is most likely to miss the deadline and how to lay out the straw in advance.” This preventively reveals bottlenecks - for example, that recording lessons in the evenings actually goes twice as slow as planned. A typical beginner mistake: ask “give me a plan for launching a course” without restrictions and receive a general list from the Internet, not tied to your evenings, budget and the fact that you are alone. A good plan is always based on concrete frameworks, and not on abstract “steps to success.”
Unlock access to submit solutions for instant AI review.