Book of the Month: Garbage In, Gospel Out

Overview

March brings us a new book of the month. We’re taking the plunge with “Garbage In, Gospel Out” by Darryl J Carlton. This book is timely for folks implementing AI solutions in their organization. Let’s look at the details. 

This book is essentially a guide to managing AI implementations. Carlton brings his experience in managing solutions to the reader, with his particular flare for efficiency and eschewing bureaucracy. This focus makes the book an excellent resource for folks who are looking at implementing a solution in their organization. The book also recognizes that “AI” isn’t just the generative AI or AI agents that have been popularized in the last couple of years, but includes other machine learning implementations and specific AI models. Ultimately, the book teaches a specific set of lifecycle skills to provide guidance to a successful implementation for the reader. What is really outstanding about the book is the end of each chapter: The reader is encouraged to use the book as a “workbook” rather than just a reference, and challenges the reader to answer specific questions and do the actual work toward achieving a result. 

Of particular interest is Carlton’s drive toward efficiency. This section should be required reading for a lot of folks. In several spots in the book, Carlton reminds us of bureaucracy indicators: 

• Meetings that generate more meetings

• Decisions that require multiple approvals

• Reports nobody reads

• Committees without clear mandates

• Reviews that don’t change anything

Carlton gives the reader direct action to take to tackle the scourge of bureaucracy, noting that “Every hour spent on bureaucracy is an hour not spent on implementation.”.  What really hammers this point home to the reader is a direct case study showing the effect of bureaucracy on the delivery of a solution. 

One of the heartening things about this book too is the very detailed references to real-life case studies of AI implementations. Carlton does an excellent job of tying the methodologies presented in the book to case studies throughout, which really brings the reader to an understanding of why each piece and decision point is important.   Carlton does this by educating the reader about the AI Technical Standard put forward by the government of Australia. The standard itself has 42 statements that describe the lifecycle of an AI project. These are continually referenced throughout the book and when case studies are explained, it will often link back to success or failure related to the statements there. 

Further, Carlton exposes the reader to the Demming PDCA (Plan-Do/Deploy-Check-Act) cycle and applies that to the context of AI Solutioning. This provides a great foundation for the reader to leverage a process for delivering a viable working product. 

To conclude, this book is a great read for those who are looking to implement an AI solution, or those looking for assistance in the oversight of AI solutions at their organization. It proves to be a timely, modern resource that will help organizations cross the finish line with AI. 

About Darryl J Carlton

After 50 years in technology, from pioneering the world’s first cloud computing company to losing everything in the dot-com bust, Darryl has learned that our greatest failures become our most valuable teachings. Darryl now helps organizations in two critical areas: implementing Responsible AI Governance before disasters strike and resolving disputes when IT projects fail. As an Expert Witness, Darryl speaks from the experience born of both project success and failure. Darryl brings hard-won insight to prevention and cure. Darryl’s neurodivergent thinking, which lets him conceive of SaaS before ‘cloud’ existed, helps him recognize patterns others miss.