IDEAS, DECISIONS & LESSONS
Notes on AI, architecture and building useful products.
279 articles
29 Mar 2026 · 5 min read
ai
Google just made Gemini Code Assist free and priced Flash-Lite at $0.25/M tokens. After 15 years building production systems, here's what this cost collapse really changes about how we build.
Read more28 Mar 2026 · 5 min read
ai
Mistral's Devstral 2 achieves 72.2% on SWE-Bench at 7x cheaper than Claude Sonnet. Here's what this means for teams building AI-powered dev tooling in 2026.
Read more28 Mar 2026 · 12 min read
ai
A comprehensive guide to the modern free-tier tech stack that lets you build, deploy, and scale a startup for roughly $20/month. No servers. No DevOps team. No funding required. Just an idea and WiFi.
Read more28 Mar 2026 · 6 min read
ai
GPT-4 level AI cost $30/M tokens in 2023. Today it's under $1. Here's the technical architecture that lets you capture 90%+ of that savings without sacrificing quality.
Read more27 Mar 2026 · 5 min read
ai
Deep dive into Gemini 3.1 Pro's practical developer features — 77.1% ARC-AGI-2, dynamic thinking API, 1M token context, and what these benchmarks actually mean when building production AI systems.
Read more27 Mar 2026 · 6 min read
agents
90% of developers now use AI at work. But the real shift in March 2026 is agents moving from suggestion-mode to autonomous execution. Here's what that actually looks like in production systems and what breaks when you go too far too fast.
Read more26 Mar 2026 · 6 min read
ai
Beyond benchmarks — a real comparison of DeepSeek V3.2 and Gemini 3 Pro for production developer workflows. Cost, code quality, context window, multimodal capabilities, and when to use which.
Read more26 Mar 2026 · 6 min read
agentic
March 2026 marks a turning point — AI coding tools evolved from assistants to autonomous agents. As a Technical Lead, here's what you actually need to know to navigate JetBrains Central, GitHub Agent HQ, Dapr Agents, MCP, and the multi-agent revolution.
Read more24 Mar 2026 · 6 min read
ai
Mistral's Devstral 2 scores 72.2% on SWE-bench, ships under MIT license, and costs up to 7x less than Claude Sonnet. Here's how it works, when to use it, and whether open-source agentic coding is finally production-ready.
Read more
