Framework
AI-Augmented Engineering
Using AI to accelerate boilerplate and pattern-matching while maintaining full human architectural ownership — every line committed is understood and owned by a developer. The opposite of vibe coding at production quality.
How it works
- 1The standard: AI handles boilerplate and speeds up pattern-matching; humans review, understand, and own every line they commit to the codebase. No exceptions.
- 2Pull request discipline: at Woz, every PR requires two developer reviews. AI-generated code goes through the same review as hand-written code. The author is responsible for what they commit.
- 3The LLM-era skill gap: engineers who came up entirely with LLMs available never faced the week-long debugging sessions that build deep architectural intuition. The easy out short-circuits deep understanding.
- 4The risk of over-reliance: for complex, high-stakes systems, blind acceptance of AI output produces technically plausible code that fails in non-obvious ways. Understanding beats velocity when correctness is required.
- 5Practical framework: use AI for speed on low-risk tasks; apply maximum human scrutiny to architecture decisions, security-sensitive code, and anything that touches customer data or payments.
Hear it on the podcast
AI That Turns Ideas Into Real CompaniesListen to the episodeMore from this episode
Build Backwards from Business RequirementsWoz's core development philosophy: architecture decisions, security guardrails, and compliance requirements are defined at the start from a CTO lens — not added iteratively after prototyping proves the concept.Distribution Before IdeaIn an era when building costs approach zero, idea originality is no longer the primary differentiating variable. Who is launching matters more than what is launching — distribution, credibility, and persona-fit are the new moats.
Get the next framework first
One email a week with the frameworks, tools, and founder stories from the podcast. Free.
More frameworks for AI & Automation
AI Copilot ArchitectureEvery AI output surfaces with a transparent reasoning chain and supporting citations before reaching a human decision-maker. If the model can't explain why, the inference doesn't get surfaced. Human judgment is preserved at the decision layer.User-First AI IntegrationStart with the exact moment of user friction, not with the AI capability. Bring AI invisibly into the user's existing behavior rather than forcing users to adopt a new interaction model to access AI features.AI-Native SDLCApply the proven software development lifecycle (epic → requirement → atomic task) to AI agent workflows. Structured decomposition puts agents on rails and produces working software instead of context-drifted hallucinations.Context Grounding for AgentsAI coding agents operating without codebase knowledge generate generic, architecturally incompatible software. Ground agents in five types of codebase documentation before generating any requirements or tasks.The Four-Category AI Coding TaxonomyAI coding tools fall into four distinct categories with different users, capabilities, and failure modes: (1) generalist code assistants (Copilot, Cursor) for developers; (2) prototyping tools (Lovable, Bolt, v0) for fast front-end visualization; (3) lifecycle agentic flows (Devin) for engineering teams across the full SDLC; (4) MVP builders (Lio) for full-stack requirements-to-scaffold generation. Knowing which category you're in tells you what to expect, where you'll hit walls, and where to go next.Intent-First DevelopmentStart with expressed intent in plain English. The planning layer converts raw founder intent into agent-ready specifications — preserving the momentum and dopamine of idea exploration while producing the precision execution requires.