jonoropeza.com
Making software and software development teams. Mostly the people parts.
Practical Lessons From Six Months Using AI All Over The SDLC - August 31st, 2026
LLM-generated code doesn’t replace the need for software engineers: hard problems remain hard, while code review and functionality verification become the new bottlenecks. AI will challenge your first principles; some need revising or refactoring, and some might need to be thrown away. Some of the added output capacity must go into care & feeding. Everything is being built like a garden being sown, a series of rapid 80/20s, which raises some questions: what kind of engineer becomes more valuable when writing code becomes the cheapest part of the SDLC, and what becomes scarce when intelligence becomes abundant?
Avoiding AI-Driven Complexity Inflation - April 26th, 2026
At this point, in late April 2026 and with 4.7 and 5.5 in the wild, agentic tooling - whether it's Claude Code, Codex, Cursor, Kiro, whatever - should be adding some level of capacity to your software development team. If it's not, start here, and if your calibration was from before about November 2025 and based on what LLM tools could do prior to Claude 4.5,
On Using AI To Make Software (in early 2026) - March 29th, 2026
A practical look at AI in software development from the perspective of a working software engineer: where LLMs and coding agents actually help across the SDLC, and where the hype outruns reality. This essay explores AI coding workflows, production software, prompt loops, code review, testing, operations, and the skills software engineers need as syntax generation gets cheaper and faster. It also examines the future of software engineering careers, junior developer hiring, bootcamps, and computer science in an era of agentic tools, automation, and fast-changing AI capabilities.
Claude 4.6 - February 25th, 2026
My initial impressions of Claude 4.6 after using it for a week to make software. This is going to change everything, both slowly and all at once.
On Building AI Applications - August 6th, 2025
Learn the core principles from 12-Factor Agents on how to design reliable LLM applications. Discover why single-prompt systems fail, how to scale beyond the easy 80%, and when to use human-in-the-loop AI. Essential insights for engineers, founders, and product leaders building AI agents in production.
Letting o1 Cook - February 8th, 2025
An exercise where I guide OpenAI's o1 through coding a JavaScript game that runs in the browser
Models of Automation - February 5th, 2023
Four of the primary models of automating any process