TL;DR
Nicola Sahar has published a free, MIT-licensed manual called 'Morphic Programming' that lays out first principles for building agentic AI systems with CLI agents such as Claude Code. The GitHub repository collects nine principles, system design advice, example commands and practical tips, and invites feedback and future updates.
What happened
A public repository on GitHub titled 'morphic-programming' hosts a free manual by Nicola Sahar that sets out first principles for creating useful AI systems driven by command-line agents. The document was framed in response to a Dec. 26, 2024 tweet from Andrej Karpathy describing a need for a practical mental model around agentic systems. The manual enumerates nine principles — including Morphability, Abstraction, Recursion, Internal Consistency, Reproducibility, Morphic Complexity, E2E Autonomy, Token Efficiency, and Mutation & Exploration — and supplements them with guidance on repo structure, git usage, context engineering, psychological tips, and example commands. The project is released under the MIT License, is available to star or watch, and lists basic repo activity such as 32 stars, two watchers and no forks or published releases.
Why it matters
- Provides a concise set of first principles aimed at helping developers adapt to agent-driven workflows Karpathy described.
- Offers practical system-design and engineering guidance for building CLI-based agent systems.
- Released under an MIT license, enabling reuse and adaptation by other practitioners and teams.
- Signals a community-oriented approach—maintainer invites issues, DMs and contributions for future parts and updates.
Key facts
- Author: Nicola Sahar, identified as founder and seller of an AI healthcare startup (Semantic Health) and now researching AI for consciousness and mental health.
- Repository: nicolasahar/morphic-programming on GitHub, containing morphic_programming_manual_v1.md and a README.
- Nine first principles listed: Morphability; Abstraction; Recursion; Internal Consistency; Reproducibility; Morphic Complexity; E2E Autonomy; Token Efficiency; Mutation & Exploration.
- Supplementary material covers system design (repo structure, git, context engineering), practical tips and example commands.
- Inspired explicitly by Andrej Karpathy's Dec. 26, 2024 tweet about the need for a manual for agentic systems.
- License: MIT — the manual is free to use, share and adapt.
- Repository activity shown on the page: 32 stars, 2 watchers, 0 forks, and no published releases as of the source snapshot.
- The README and manual indicate plans for more principles, advanced topics, and a compressed version for system context in future updates.
What to watch next
- Planned follow-ups listed in the repo: additional principles, advanced topics and a compressed version for embedding as system context.
- Repo activity and contributions — the author asks readers to star/watch and open issues or send DMs with feedback.
- No formal releases have been published yet (repo shows no releases); watch the repository for updates.
Quick glossary
- Agentic AI: AI systems designed to perform tasks autonomously or semi-autonomously, often by executing sequences of actions or commands.
- CLI agent: A command-line interface program that operates as an agent, accepting text commands and producing outputs or triggering tools.
- Reproducibility: Design and engineering practices that allow a system's behavior to be reliably recreated after failures or over time.
- Token efficiency: The practice of maximizing useful work per input token when interacting with language models to reduce cost and latency.
- Morphability: Treating natural language prompts as a malleable form of code that can be shaped to express commands, abstractions, and workflows.
Reader FAQ
Is the manual free to use?
Yes. The project is published under an MIT License and hosted publicly on GitHub.
Which platforms or agents does the manual target?
The manual names CLI agents like Claude Code as an example; broader platform applicability is not elaborated beyond that.
How can readers give feedback or suggest additions?
The author invites readers to open issues on the repository or send direct messages as noted in the repo.
Are there published releases or a versioned roadmap?
Not confirmed in the source whether a formal release schedule or versioned roadmap exists; the repo shows no published releases and lists planned future content.
Morphic Programming: A First Principles Manual for Agentic AI By Nicola Sahar What is this? A free manual on first principles for building useful AI systems with CLI agents like…
Sources
- I wrote the manual Karpathy said was missing for agentic AI
- Sahar's Coding with AI guide
- Semantic Health & the Future of Medical Coding with Dr. …
- How We're Unlocking Data-Driven Care Applications By …
Related posts
- Ask HN: What tech jobs or niches allow a developer to do minimal real work?
- Microsoft CEO Satya Nadella pushes new metaphor: AI as a lever, not job killer
- Grok Can’t Apologize — Why Headlines Say the Bot Took the Blame