Vibe Engineering Effect Apps — Michael Arnaldi, Effectful
TL;DR
Michael Arnaldi demonstrates "vibe engineering" by building a TypeScript project with AI agents, revealing that cloning library repositories directly into your codebase—rather than using npm packages—enables AI to learn patterns from source code, while strict TypeScript and custom lint rules act as essential guardrails.
🧠 LLM Architecture Constraints 2 insights
Context windows act as fixed memory arrays
Unlike human brains, LLMs cannot form long-term memories from conversations, operating solely within limited token windows that cause遗忘 of previous instructions.
Post-training freezes knowledge updates
Models receive no continuous learning after deployment, meaning they lack awareness of libraries released after their training cutoff dates.
📁 Repository Cloning Strategy 2 insights
Clone dependencies into project directories
Placing library source code directly in your repo (e.g., `repos/effect`) allows AI agents to analyze implementation patterns, whereas node_modules and gitignored files remain invisible to coding agents.
Source code supersedes documentation
AI agents trained on codebases perform better reading actual TypeScript implementations than parsing human-written documentation or using MCP servers.
🛡️ Safety Guardrails for AI Coding 2 insights
Maximize TypeScript strictness
Configure all diagnostics as errors and enable format-on-save to force AI agents to fix type issues immediately rather than accumulating technical debt.
Implement custom ESLint rules
Ban explicit type assertions (`as X`) and `any`/`unknown` types to prevent AI from circumventing Effect's type safety when it encounters complex type machinery.
⚙️ Development Workflow Setup 2 insights
Create agents.md instruction files
Maintain a living document listing available commands (type check, test) and repository locations to guide AI behavior without context pollution.
Use Effect v4 for agent safety
Effect's structured error handling and type safety prevent AI-generated code from creating unmaintainable "small" snippets lacking proper error management.
Bottom Line
Clone library repositories directly into your project as subtree sources rather than npm dependencies, while enforcing strict TypeScript and custom ESLint rules to constrain AI agents to type-safe patterns.
More from AI Engineer
View all
The Production AI Playbook: Deploying Agents at Enterprise Scale — Sandipan Bhaumik, Databricks
Sandipan Bhaumik from Databricks presents a battle-tested five-pillar framework for deploying enterprise AI agents, arguing that starting with model selection leads to inevitable production failures while proper evaluation, observability, and data governance determine success at scale.
Sovereign Escape Velocity: Ownership w Open Models — Gus Martins, & Ian Ballantyne, Google DeepMind
Google DeepMind's Gus Martins and Ian Ballantyne introduce Gemma 4, a family of open models (2B to 31B parameters) that deliver frontier-level intelligence with disproportionate efficiency, enabling sovereign AI ownership through local deployment, Apache 2.0 licensing, and on-device capabilities.
LLM Observability, Evaluation, Experimentation Platform — Dat Ngo, Arize
Dat Ngo from Arize AI explains how modern AI systems require reimagined observability and evaluation patterns built on OpenTelemetry to manage non-deterministic agents, emphasizing that the future of AI engineering lies in automated experimentation flywheels that eliminate manual dashboard work.
Text Diffusion — Brendon Dillon, Google DeepMind
Google DeepMind researcher Brendon Dillon explains text diffusion as a parallel alternative to autoregressive language models that iteratively denoises random tokens rather than generating sequentially, offering significantly lower latency and unique capabilities like self-correction and adaptive computation, though currently limited by high serving costs for large batches.