The date(created_at) example is the sharpest one, a query that filters to a single day but silently loses partition pruning because the filter function isn't sargable is exactly the kind of thing invisible from reading the SQL, but obvious the moment you see a full-table scan in the plan.
The Explain vs. Profile distinction is the right one to draw clearly, a general chatbot can describe what your SQL does, but it has no way to know your table is 3TB or that the row estimate was off by four orders of magnitude. That data only exists in the warehouse, which is the actual value-add here over pasting SQL into any LLM.
The date(created_at) example is the sharpest one, a query that filters to a single day but silently loses partition pruning because the filter function isn't sargable is exactly the kind of thing invisible from reading the SQL, but obvious the moment you see a full-table scan in the plan.
The Explain vs. Profile distinction is the right one to draw clearly, a general chatbot can describe what your SQL does, but it has no way to know your table is 3TB or that the row estimate was off by four orders of magnitude. That data only exists in the warehouse, which is the actual value-add here over pasting SQL into any LLM.