IDEAS, DECISIONS & LESSONS
Notes on AI, architecture and building useful products.
919 articles
10 Jun 2026 · 8 min read
architecture
Daily English practice for tech professionals. Evening session — review architecture vocabulary, learn to explain complex systems in plain English, and practice speaking with technical stakeholders.
Read more10 Jun 2026 · 5 min read
architecture
Daily English practice for tech professionals. Morning session covering architecture vocabulary — microservices, distributed systems, cloud-native design with pronunciation and real-world examples.
Read more10 Jun 2026 · 8 min read
architecture
Daily English practice for tech professionals. Noon session — architecture vocabulary deep dive: scalability, latency, microservices, and how to explain system design to non-tech stakeholders.
Read more10 Jun 2026 · 5 min read
ai
Google's Gemini 3.5 Flash hit general availability with 4x faster inference than competing frontier models, a 1M token context window, and $1.50/$9 pricing. It outperforms Gemini 3.1 Pro on coding and agentic benchmarks. Here's the honest assessment of what this means.
Read more10 Jun 2026 · 8 min read
connected-speech
Native English speakers link, reduce, and blend words in ways no textbook teaches. If you pause between every word, you sound unnatural even with perfect grammar. Here are the four patterns of connected speech with tech-specific scripts to practice.
Read more10 Jun 2026 · 8 min read
english
Running a meeting in your second language is harder than attending one. You have to open, steer, redirect, and close — all in real time. Here's the language structure for each phase, from a Vietnamese tech lead's perspective.
Read more10 Jun 2026 · 5 min read
ai
Google's TurboQuant research reduces KV cache memory overhead by ~100x using a two-step algorithm combining PolarQuant vector rotation and Johnson-Lindenstrauss compression. This could make 2M-token context models economically feasible for many more teams.
Read more10 Jun 2026 · 8 min read
code review
Code review comments in English have three failure modes: too vague to act on, too harsh to receive well, or too polite to land at all. Here's the language for reviews that teach, not just judge — from how to open a review to how to handle pushback.
Read more09 Jun 2026 · 5 min read
ai
Anthropic documented how DeepSeek, Moonshot, and MiniMax ran 16 million fraudulent Claude queries through 24,000 fake accounts to distill Claude's reasoning, tool use, and coding capabilities into their own models. This is what the attack looked like, and why it matters.
Read more
