You might be confusing this with "Productizing Twilio Applications", which was about the engineering challenges of moving from Twilio code samples to reliable ways to test, architecture, and deploy Twilio applications. This is rather unlike the current email, which was written today and is (to the extent I'm capable of talking about new things after eight years of writing a metric effton about this industry) all new, and focused on how to get consulting companies to turn their buy-our-time-in-discrete-chunks model into a productized buy-our-value-without-necessarily-our-time model.