3 points | by josephinepqt a day ago ago
3 comments
Hybrid is the answer. Local for high-volume/low-stakes, APIs for quality-critical tasks.
Token costs matter at scale. Engineering time matters more when you're small.
It's too slow for agentic flows
can you give an example?
Hybrid is the answer. Local for high-volume/low-stakes, APIs for quality-critical tasks.
Token costs matter at scale. Engineering time matters more when you're small.
It's too slow for agentic flows
can you give an example?