The AI Engineering Notebook That Outgrew Prompt Hype

What it does

This repository is a curated knowledge base of markdown files tracking the state of generative AI and large language models. Maintained by the author of the latent.space newsletter, it collects links, use cases, reading lists, and rough notes across text, image, audio, code, and infrastructure. The Resources folder holds cleaned-up canonical references, while files like TEXT.md, IMAGE_GEN.md, and INFRA.md act as living scratchpads for specific domains.

The interesting bit

The repo was originally named prompt-eng but was deliberately renamed after its owner decided prompt engineering was overhyped, pivoting to the broader “AI Engineering” umbrella. That arc—from narrow tactic to sprawling field—mirrors how many developers have had to recalibrate their own mental models over the last few years. It is less a finished product than a public Zettelkasten for the AI gold rush.

Key highlights

• Covers the full generative AI stack: text generation, ChatGPT competitors, semantic search, code models, audio/music, image generation (heavy on Stable Diffusion), and emerging agentic AI.
• Includes curated reading lists tiered by difficulty, from beginner explainers to intermediate research recaps and annual State of AI reports.
• Tracks concrete use cases—synthetic MRI scans, text-to-3D, game asset generation, and AI music videos—rather than just abstract capabilities.
• The Resources directory offers permalink-ready references distinct from the raw, frequently updated note files.
• 6,233 stars suggest it has become a de facto bookmark folder for a lot of developers.

Caveats

• The README itself is only a high-level overview; the actual content is scattered across a dozen other markdown files, so navigation is utilitarian at best.
• Many sections are explicitly labeled “raw notes” or “stub notes,” meaning depth and polish vary significantly by topic.
• Heavy reliance on external links (especially Twitter/X posts) means rot is inevitable; the value is in the curation, not permanence.

Verdict

Worth a star if you are a software engineer drowning in AI news and want a human-filtered signal. Skip it if you are looking for a software library, runnable code, or a structured course—this is a reading list and notebook, not a tool.