Personal blog on ML, AI agents, recommendation systems, and GPU programming.https://louiswang524.github.io/https://louiswang524.github.io/blog/agent-topology-decision-guide/https://louiswang524.github.io/blog/agent-topology-decision-guide/Five multi-agent design patterns — with a decision guide for when each one earns its complexity and when it will burn you.Tue, 19 May 2026 00:00:00 GMThttps://louiswang524.github.io/blog/ssa-subquadratic-sparse-attention/https://louiswang524.github.io/blog/ssa-subquadratic-sparse-attention/Why standard attention breaks at 128K tokens, how four families of efficient attention tried and partially failed to fix it, and how content-dependent sparse routing achieves linear scaling without sacrificing retrieval accuracy.Fri, 15 May 2026 00:00:00 GMThttps://louiswang524.github.io/blog/agent-harness-poker/https://louiswang524.github.io/blog/agent-harness-poker/How a Texas Hold'em simulator became a blueprint for any domain where autonomous agents compete, negotiate, and adapt — turn by turn.Sat, 25 Apr 2026 00:00:00 GMThttps://louiswang524.github.io/blog/llm-knowledge-base/https://louiswang524.github.io/blog/llm-knowledge-base/Inspired by Andrej Karpathy's post on LLM knowledge bases, I built a system where Claude Code skills manage a personal wiki end-to-end — ingesting raw content, compiling concept articles, synthesizing connections, and answering questions. You never touch the wiki. The LLM owns it.Sun, 05 Apr 2026 00:00:00 GMThttps://louiswang524.github.io/blog/gemma-family/https://louiswang524.github.io/blog/gemma-family/A technical deep-dive into Gemma 4's four core ideas — MatFormer elastic inference, hybrid attention with p-RoPE, parallel dense+MoE FFN, and native agentic tooling — with the Gemma 1–3 lineage as context.Thu, 02 Apr 2026 00:00:00 GMThttps://louiswang524.github.io/blog/turboquant-explained/https://louiswang524.github.io/blog/turboquant-explained/A technical breakdown of Google Research's TurboQuant stack: why KV-cache quantization is really an inner-product estimation problem, how PolarQuant removes normalization overhead, and where QJL fits into the final system.Thu, 02 Apr 2026 00:00:00 GMThttps://louiswang524.github.io/blog/diffusion-llm/https://louiswang524.github.io/blog/diffusion-llm/A technical deep-dive into continuous and masked diffusion LLMs — full derivations, key models (LLaDA, Dream, Mercury), head-to-head comparison with autoregressive LLMs, and an honest look at whether dLLMs can replace AR in the future.Wed, 01 Apr 2026 00:00:00 GMThttps://louiswang524.github.io/blog/claude-code-multi-agent-patterns/https://louiswang524.github.io/blog/claude-code-multi-agent-patterns/Claude Code can run multiple Claude instances in parallel within the same process. Here's how in-process teammates, permission sync, and coordinator mode work — and what it means for building your own multi-agent system.Tue, 31 Mar 2026 00:00:00 GMThttps://louiswang524.github.io/blog/claude-code-plugin-skill-system/https://louiswang524.github.io/blog/claude-code-plugin-skill-system/Claude Code has two extension points: skills (slash commands that inject prompt content) and plugins (packages that contribute commands, MCP servers, and hooks). Here's how each works and why they're designed differently.Tue, 31 Mar 2026 00:00:00 GMThttps://louiswang524.github.io/blog/claude-code-security-permissions/https://louiswang524.github.io/blog/claude-code-security-permissions/Claude Code runs shell commands, edits files, and makes network requests on your machine — without a kernel sandbox. Here's the permission model that makes this safe enough to ship.Tue, 31 Mar 2026 00:00:00 GMThttps://louiswang524.github.io/blog/claude-code-tool-use-loop/https://louiswang524.github.io/blog/claude-code-tool-use-loop/A tool call is a structured JSON request from the LLM to run a named function. Here's exactly how Claude Code handles the full lifecycle — from API call to file edit to loop continuation.Tue, 31 Mar 2026 00:00:00 GMThttps://louiswang524.github.io/blog/demystifying-claude-code-architecture/https://louiswang524.github.io/blog/demystifying-claude-code-architecture/Claude Code is one of the most widely used AI coding tools and one of the least understood. We read the source so you don't have to.Tue, 31 Mar 2026 00:00:00 GMThttps://louiswang524.github.io/blog/genrec-paradigm-comparison/https://louiswang524.github.io/blog/genrec-paradigm-comparison/A head-to-head comparison of the two paradigms remaking recommendation — semantic ID autoregressive models and fine-tuned LLMs — with trade-off analysis and a look at how they're converging.Sun, 29 Mar 2026 00:00:00 GMThttps://louiswang524.github.io/blog/attention-variants/https://louiswang524.github.io/blog/attention-variants/From vanilla multi-head attention to Flash Attention 3 — the engineering bottlenecks that drove every major attention variant and the math behind each fix.Sat, 28 Mar 2026 00:00:00 GMThttps://louiswang524.github.io/blog/harness-is-the-moat/https://louiswang524.github.io/blog/harness-is-the-moat/Model quality is commoditising. The durable competitive advantage in 2026 is harness architecture — the deterministic enclosures that make probabilistic agents reliable. A deep analysis of the four architectural primitives every production harness must implement, and how Autoresearch, Ralph Loop, Superpowers, and GSD each solve them differently.Wed, 25 Mar 2026 00:00:00 GMThttps://louiswang524.github.io/blog/generative-retrieval/https://louiswang524.github.io/blog/generative-retrieval/From semantic IDs to OneRec Think — how Meta, Kuaishou, Google, Alibaba, ByteDance, and LinkedIn are replacing two-stage retrieval pipelines with generative models. What's in production and where the field is heading.Mon, 23 Mar 2026 00:00:00 GMThttps://louiswang524.github.io/blog/vibe-coding-to-harness-engineering/https://louiswang524.github.io/blog/vibe-coding-to-harness-engineering/Vibe coding gets you a working prototype in 10 minutes. Harness engineering is how you ship it to production. Here's the difference, why it matters, and how to make the transition.Mon, 23 Mar 2026 00:00:00 GMThttps://louiswang524.github.io/blog/inference-cost-trends/https://louiswang524.github.io/blog/inference-cost-trends/An analysis of hardware trends, algorithmic improvements, and market forces driving down the cost of running large language models.Sun, 01 Mar 2026 00:00:00 GMT