The SaaS Companies That Win Will Be Agent-Ready →
Codex, Claude, and other AI agents are going to put pressure on traditional SaaS. The companies that survive will expose their workflows, reduce friction, and become tools agents can actually use.
Thoughts on software engineering, AI, and building things. Posts are written with AI assistance.
Codex, Claude, and other AI agents are going to put pressure on traditional SaaS. The companies that survive will expose their workflows, reduce friction, and become tools agents can actually use.
A childhood NFL Street memory, how it led me toward software engineering, and why AI could change how kids learn.
OpenAI and Anthropic are moving from selling AI tools to financing and staffing AI deployment companies. That shift will create jobs, displace work, and make the next phase move fast.
A small Simplified Chinese prompt made me think harder about how AI can reduce language barriers and make even tiny software feel global sooner.
AI may push websites away from noisy interfaces and back toward clear stories, useful text, and proof that both humans and LLMs can understand.
Voice-first agents are changing work from clicking through tools to routing intent across apps, repos, files, and context.
AI will make it easier than ever to launch software. The products that last will come from people who actually have the problem and care enough to solve it well.
The four-question filter I use to choose side projects, avoid spreading myself too thin, and focus on problems worth staying with.
AI labels are a useful start, but social feeds need to change how they rank, contextualize, and amplify synthetic content.
A practical field report on using Codex and GPT-5.5 to turn a rough product idea into a working, verified AI app.
A practical starting point for system-minded people who want to build with AI, even if they do not know how to code.
My actual AI writing workflow: raw thoughts first, AI-assisted structure second, and human judgment all the way through.
How I turned a long Codex session into version-controlled instructions for how I work, write, ship, and make decisions.
The practical reasons I moved my day-to-day AI agent workflow from Claude Code to Codex, and why the hardest part was not the model choice.
Why Codex feels like it jumped versions even without a clean version number, and why capability shifts matter more than release labels.
How I use Codex's native profiles plus a tiny zsh wrapper to choose a sandbox and approval posture before the agent starts.
I used Claude Code to build a content marketing system from scratch — brand strategy, content calendars, and a reusable framework for managing multiple clients.
How I built an interactive visualization of the full MCP pipeline — tool discovery, selection, JSON-RPC execution, and response generation — in an afternoon with Claude Code.
How I built an interactive visualization of the full RAG pipeline — from document chunking to vector search to augmented generation — in an afternoon with Claude Code.
How I used Claude Code to overhaul my entire professional brand — resume, LinkedIn, personal website, and blog — from a single source of truth.
Something shifted. Everything I went to school for is changing, and I'm trying to make sense of it.
I turn problems into prototypes.