Page 3 of 14: older articles, tutorials, and videos.

Reliable automations beat fancy AI agents, and I show how to build them by maximizing deterministic code and using LLMs only where real reasoning is required. Most people asking for an AI agent do not need one: LLMs…

Automation isn't all-or-nothing: automating 80% of a task is far better than automating none of it. I argue that every task has routine, repetitive elements you can automate with traditional code, and you can use LLM…

I get reliable, fast, cheap automations by describing the process at the most granular level possible before building anything. I show how my auto-editor started as tiny chained automations — audio normalization and…

I argue you should never automate a task until you’ve standardized it with written SOPs and done it manually at least once or twice, because if you can’t do it by hand you won’t be able to automate it reliably. I show…

I show how to set up autonomous AI agents that read your customer reviews, support tickets and social media comments and turn them into short, actionable reports, so you never have to read hundreds of messages…

Gemini is a far more capable coding model than its reputation suggests, and I make the case for using it on complex projects even though OpenAI and Anthropic lead the frontier. I point to a fully automated video editor…

OpenAI serves 800 million ChatGPT users off a single Postgres write instance, and the playbook their engineers built around it is the most useful scaling material I've read as a backend developer. I walk through every…

I show how to build a free auto-captioning pipeline that generates subtitles with faster-whisper and burns them into your videos with FFmpeg — no manual captioning, no paid services, no expensive LLMs. Captions make…

I show non-technical business owners and vibe coders how to automate posting to YouTube, Instagram, TikTok, X, LinkedIn and other platforms for free using each platform’s API and an AI-generated setup. You copy my long…

OpenAI's GPT-5.6 Sol hacked Hugging Face during a cybersecurity benchmark—not because of a malicious prompt, but by finding a zero-day, escaping its sandbox, gaining internet access, escalating privileges, stealing…

Most businesses overspend on LLM tokens because they route every request to a frontier model, so I show how a model router cuts AI agent costs by up to 80%. I walk through a four-step build: sample at least 500 real…

I explain how to actually automate your business with AI by shifting from role-based thinking to workflow-based thinking: instead of trying to automate a whole job like social media manager or video editor, break it…


Dude, you’re at the bottom of my landing page.
Just book a call already if you’re that interested.
You scrolled all the way here.