What is the difference between a generator and a coroutine?

Question presented to candidate: "Generators and coroutines both pause and resume, and they even share methods like send(). So what really is the difference between a generator and a coroutine in Python, and why can you not just await a generator?"

What a strong answer should cover:

• Purpose: a generator (def + yield) produces a sequence of values for its caller; a native coroutine (async def) waits for results and is driven by an event loop.
• Who drives it: a generator is pulled with next() / for; a coroutine is resumed with send(None) by an event loop and is not an iterator (next(coro) is a TypeError).
• The shared machinery: both are suspended frames with send / throw / close, both report their return value through StopIteration.value; the code-object flags differ (GENERATOR vs COROUTINE), and inspect can tell them apart.
• Awaitability: only coroutines, Futures/Tasks, objects with __await__, and types.coroutine-decorated generators can be awaited; await on a bare generator is a TypeError.
• History: generator-based coroutines (PEP 342 send, PEP 380 yield from, @asyncio.coroutine) led to async/await in PEP 492 (3.5); @asyncio.coroutine is gone in 3.11.
• The bottom of every await chain is a yield to the event loop — "a coroutine is a generator with a different job and a stricter type".

Code / implementation expected: Optional — showing next() on a generator vs send(None) on a coroutine, or a toy event loop.