Heading home after a solid trip to Texas. Glad to be done this stretch of travel, and taking off a few days to try and recharge. No reading list next week!
[article] Understanding Spec-Driven-Development: Kiro, spec-kit, and Tessl. There seem like a dozen ways to do everything in the AI world. Here’s a look at a few options for those creating agent-friendly specs.
[paper] A Survey of Vibe Coding with Large Language Models. Good look at the state of vibe coding and what’s still true, regardless of how fast you can crank out code.
[blog] Claude Skills are awesome, maybe a bigger deal than MCP. Super interesting! We’re also drowning in techniques now (Skills, MCP, tools, agents, sub-agents, etc), but that’ll happen until we hit a logical consolidation point.
[blog] What is Configuration as Data? Brian was an early pioneer of this approach which still isn’t mainstream. But super relevant and maybe even easier to deliver now.
[blog] The ultimate prompting guide for Veo 3.1. You don’t need to be a train cinematographer to use a video generation model like Veo. Great advice here.
[article] The Best CEOs Can Spot Where Their Organization Is Under the Greatest Pressure. I liked the “three fronts” that leaders have to always protect.
[blog] Save costs and decrease latency while using Gemini with Vertex AI context caching. You don’t care about this when prototyping or working with small apps. But as soon as you start dealing with large amounts of context, you’ll want a way to reuse what you have and keep cost down.
[article] Deep Agents. Seems like a topic of conversation lately. We’re already moving past the first stage of simplistic agents to something more.
[article] The “AI is a Bubble” Narrative is Stupid, Wrong, and Dangerous. I’ll admit that I’ve found some of the recent AI funding decisions to be curious. But this is a sober take into what’s actually going on.
Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:
Richard Seroter's Architecture Musings 
Leave a comment