Judgment
Judgment is the process or capacity to assess information and circumstances, interpret them in context, and form an evaluation or decision.
Vibe Engineering
Human judgment is the reasoning and decision-making people apply when evaluating situations, evidence, risks, and software changes.
- How it is used
- Frames human judgment as the responsibility for deciding what to build and assessing the consequences of AI-generated changes.
- How deep it goes
- Distinguishes generation from judgment and develops practical evaluation questions, decision points, and guardrails rather than a formal decision procedure.
- Examples
- Uses a game initialization question and a multi-agent browser project to illustrate technique selection and the amplification of instructions.
- What you can do after
- Evaluate generated code by examining failure modes, test relevance, and the distinction between verified behavior and assumptions.
Architecting for Autonomy
A reasoned determination that selects or evaluates a course of action under conditions requiring interpretation rather than mere rule execution.
- How it is used
- Structures high-stakes value streams around explicit decision points, delegated authority, confidence bands, escalation, and accountability.
- How deep it goes
- Maps decision surfaces and governance fields, but stops short of prescribing domain-specific policies or replacing professional judgment.
- Examples
- Uses retail banking decisions, secured lending origination, KYC, fraud thresholds, vulnerability assessment, and legal knowledge graphs.
- What you can do after
- Apply it by mapping where ambiguous reasoning occurs and assigning bounded authority, escalation triggers, policy references, and accountable roles.
Introduction to Generative AI, Second Edition
Human judgment is the process by which people assess information and make decisions using experience, reasoning, and contextual understanding.
- How it is used
- Guides human oversight, responsible AI adoption, and decisions about when model outputs require review.
- How deep it goes
- Connects judgment to agency, accountability, literacy, and error detection but stops short of a formal decision procedure.
- Examples
- Uses error checking, output revision, citations, confidence scores, and revision histories as aids to informed judgment.
- What you can do after
- Evaluate AI outputs critically, revise them, and decide when to involve human judgment.
Read how Introduction to Generative AI, Second Edition explains judgment →
Also covered in
- AI Model EvaluationIn early access (MEAP)
- Evaluating AI SystemsIn early access (MEAP)
- Build High Value Software SystemsIn early access (MEAP)
- Interview Speak
- Demand Forecasting Best Practices
What to know first
- language modelA language model is a computational model that represents patterns in linguistic sequences to assign probabilities, predict, generate, or evaluate language.
- artificial intelligenceArtificial intelligence is the field of computing concerned with creating systems that perform tasks associated with intelligence, such as learning, reasoning, perception, and decision-making.
Related concepts
- LLM-as-a-judgecontrasts with · LLM-as-a-judge is an evaluation technique in which a language model assesses or compares outputs according to specified criteria.
- code reviewused by · Code review is the systematic examination of source code or proposed changes to assess correctness, quality, risks, or conformance to requirements or standards.
Questions about judgment
What is judgment?
Judgment is the process or capacity to assess information and circumstances, interpret them in context, and form an evaluation or decision.
Which Manning books explain judgment?
Vibe Engineering, Architecting for Autonomy and Introduction to Generative AI, Second Edition.
What should I know before learning judgment?
Manning books that teach judgment usually introduce language model and artificial intelligence first.
What is the difference between judgment and LLM-as-a-judge?
Judgment is the process or capacity to assess information and circumstances, interpret them in context, and form an evaluation or decision. LLM-as-a-judge is an evaluation technique in which a language model assesses or compares outputs according to specified criteria.
What kinds of judgment are there?
Kinds of judgment include engineering judgment, grammatical acceptability judgment and philosophical judgment.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.


