Nam
Posts
All the articles I've posted.
-
Building AceSpeak: The Dream Is Almost Here
From June to August 2026, we turned AceSpeak from a shared belief into a real beta. A co-founder and CIO’s note from the final stretch before launch.
-
Kimi K3 for SMEs: Pilot Now, Host Later
A practical follow-up on Kimi K3's open-model approach: what is actually available, why a 2.8T model is not an office-server deployment, and how SMEs can evaluate API, managed and private options.
-
Loop engineering for SEO: improve one page at a time
A practical SEO loop that defines a measurable outcome, improves one page, reads the evidence and uses it to repeat, revert or stop.
-
Kimi K3 May Win Asia Without Beating GPT or Claude
Moonshot's Kimi K3 does not need the benchmark crown to change Asia's AI market. Its real advantage is the combination of frontier capability, regional product fit, open weights, and a much lower frontier price.
-
AI agent loop engineering: what it is and how to use it
Loop engineering moves the hard work from repeatedly prompting an AI agent to designing the goal, tools, evidence, memory, limits and stopping rules around it. Here is a practical way to start.
-
How to make your AI agent less forgetful: a beginner's guide
AI memory is not magic. This beginner-friendly guide shows how preferences, project spaces, a one-page notebook, handovers, and careful retrieval make an AI agent feel far less memoryless.
-
Five prompt habits for getting everyday work done with AI agents
You do not need a technical background to delegate well to an AI agent. These five prompt habits turn vague daily requests into useful, safe, and reviewable work.
-
Five prompt habits that work in both Claude Code and Codex
The best coding-agent prompts are not long incantations. These five practical habits—drawn from the official Claude Code and Codex guidance—make implementation faster, safer, and easier to verify.