4 years ago

Category: Deployment, Infrastructure, machine-learning

Running inference, even at small scale, is expensive. Justifying this cost sucks, particularly when people are being able to serve 10,000 users a day on a typical CRUD app using just AWS’s free tier.