About HelloAI
An information-leading, fun, systematic AI learning and news portal.
70
Articles (中文)
28
Paper deep-dives
15
Visualizations
31
Articles (EN)
What This Is
HelloAI is an entry-point AI site for Chinese readers at all levels. We believe AI shouldn't be the toy of a select few— whether you're a complete beginner, already using ChatGPT daily, or writing LLM applications, there's something here for you.
Three No's
- No fluff — we don't say "AI will change the world"; we tell you exactly what works and what doesn't.
- No jargon flexing — every term is explained the first time it appears.
- No blind hype — when a new paper / model drops, we explain what it is, why it matters, what its limits are before deciding if it's worth your time.
Four Sections
- News — latest AI developments, auto-fetched daily + curated.
- Learning paths — L0-L7 eight levels, from AI 101 to distributed training, systematic.
- Visualizations — abstract concepts as toys you can play with: Tokenizer, Attention, Diffusion, KV Cache...
- Papers — deep-dives on classic and contemporary must-read papers.
L0 → L7 Learning Path
The most distinctive feature. Eight levels by careful design:
🥚
L0 AI for Everyone
No technical background needed
🐣
L1 Math & Python
Calculus, linear algebra, probability, NumPy, PyTorch
🐣
L2 Classical ML
Regression, classification, clustering, trees, ensembles
🐥
L3 Deep Learning Core
MLP, CNN, RNN, Attention, Transformer
🐥
L4 LLMs & Generative AI
Tokenizer, training, RAG, Agents, MCP
🐥
L5 Multimodal & Frontier
Diffusion, ViT, CLIP, Sora, Whisper
🐥
L6 AI Safety & Alignment
RLHF, CAI, interpretability, policy
🐥
L7 Systems & Engineering
GPU, distributed training, inference, monitoring
Tech Stack
- Astro 4 — content-driven static site generator, native MDX + i18n + file routing
- Tailwind CSS — styling
- React Islands — for interactive visualizations (lazy-loaded)
- Pagefind — static full-text search (no backend)
- Giscus — comments via GitHub Discussions
- nginx + VPS — deployment, GitHub Actions auto-sync
Open Source & Content License
Site code: MIT. Content: CC BY-NC-SA 4.0—free to republish with attribution, non-commercial, same license. Contributions welcome.
In One Sentence
We're building this so any reader can stay informed about AI—not held back by information gaps— whether you want to learn, or just want to know what's happening.