問題文
A five-person team has no dedicated hardware and needs a working assistant within two weeks. Usage is expected to swing between a handful of requests per day and several thousand during month-end. Which hosting choice matches this situation, and why?
選択肢
- Waiting until the traffic pattern is stable, because a hosting decision cannot be revisited later.
- Running a model on each member's laptop, because that removes any dependency on a network and gives the same throughput as a server.
- Buying two accelerator servers now, because owning hardware is always cheaper per request than paying a provider.
- A cloud-hosted model accessed through an API, because capacity is the provider's concern and the team pays for the tokens it actually uses.