llm_ai_technology_notes

LLM and AI Technology Notes

Living notes on LLM technology, recent progress, infrastructure, and the surrounding AI ecosystem. The repo currently uses standalone HTML pages for long-form explainers, technical notes, and visual summaries.

Start Here

• AI startup landscape by supercycle layer
• Economics of the AI supercycle
• Inference technology notes

Pages

File	Description
ai-supercycle.html	Long-form notes on the economics of the AI supercycle, including value accrual by stack layer, inference economics, power/data-center buildout, and where durable value may migrate.
ai_layers_slide.html	A one-page visual landscape of private AI startups by supercycle layer, with reported valuation signals through July 2026.
inference-technology.html	Technical notes on inference architecture and related AI infrastructure.

Viewing

Open any HTML file directly in a browser:

open ai_layers_slide.html

If GitHub Pages is enabled for this repository, the pages are available at:

https://hefeicoder.github.io/llm_ai_technology_notes/ai-supercycle.html
https://hefeicoder.github.io/llm_ai_technology_notes/ai_layers_slide.html
https://hefeicoder.github.io/llm_ai_technology_notes/inference-technology.html

GitHub Pages can take a few minutes to build after a push. If a page returns 404 immediately after enabling Pages or pushing changes, wait for the Pages build to finish.

Notes on Data

The startup landscape slide mixes several types of valuation signals:

• completed private funding rounds
• reported private-market targets
• IPO targets
• acquisition or strategic transaction values
• undisclosed rounds where only raise size is public

Cards marked with * are not clean post-money funding valuations. Re-check fast-moving valuation data before publishing or quoting externally.

Editing

The files are plain HTML/CSS with no build step. Keep updates self-contained unless a shared style or generated pipeline is introduced later.