Almost Right: How AI Is Quietly Removing the People Who Check Its Work is a completed adult nonfiction manuscript of approximately 66,000 words.
Almost Right argues that the central problem created by generative AI is not simply hallucination. It is the verification gap: AI can produce plausible work at a rate that exceeds our ability to independently judge it, while the same automation can weaken the apprenticeship and repeated practice through which future experts learn to make those judgments.
The book moves from the history of AI and its commercial incentives into software security, medicine, law, education and labor markets, then closes with a practical framework for preserving independent verification rather than rejecting the technology.
The public conversation tends to split between acceleration and catastrophe. This book takes a narrower, testable position: the output is becoming abundant faster than expert judgment is. That matters whether AI ultimately creates more jobs or fewer, whether model capability plateaus or accelerates, and whether regulation tightens or loosens.
Vila is an AI user and builder writing from outside the research establishment. The narrative begins with the practical experience of using AI to cross expertise boundaries and discovering that producing a result and verifying the result are separate forms of labor.
Standard submission-format DOCX.
Download →Publication-focused verification work supporting the manuscript.
Review →Planned charts, figures and documentary images.
Review →This page is intended as a supplementary project hub. It does not replace an agency's stated submission procedure. Queries, proposals, sample chapters and attachments should still be submitted exactly as each agent requests.
The formal query still needs to be tailored to each agent and submitted through that agent's required channel. These files are the underlying package.
Overview, audience, comps, author/platform, marketing plan and all 29 chapter summaries.
Download proposal →Chapters 1–3 in clean submission format.
Download sample →Co-Intelligence demonstrates demand for practical, evidence-based guidance on living and working with AI. The AI Con demonstrates demand for accessible skepticism toward AI hype. Empire of AI demonstrates trade appetite for consequential reporting about the institutions driving the boom.
Almost Right occupies a different lane: it is neither an adoption manual nor an anti-AI polemic. It argues that generation is becoming abundant faster than independent verification—and that automation can weaken the apprenticeship system that creates tomorrow's verifiers.
The platform is developing rather than inherited from a large preexisting following. Current assets include this complete September reader, the industry portal, a publication-focused claims audit, source/rights documentation, and a finished manuscript designed for interviews, essays, podcasts and speaking. The marketing plan targets AI, business, cybersecurity, education, future-of-work and professional audiences without representing prospective reach as an existing audience.