Skip to content

AI Techbom Lab

Real builds. Real experiments.

The AI Techbom Lab is where we build, test and break things in public — then write down what actually worked. No hype, no re-hashed news: real builds, demos, comparisons and experiments.

lab / what the lab covers8 tracks
  • T01

    AI Agents

    Building and stress-testing agent workflows to find out where they hold up and where they fall over.

  • T02

    AI Coding

    Hands-on comparisons of AI coding tools and workflows on real codebases, not toy demos.

  • T03

    Next.js & TypeScript

    Practical patterns for fast, accessible, type-safe web apps — App Router, data, performance and SEO.

  • T04

    Software Engineering

    Architecture decisions, trade-offs and refactors explained through real projects.

  • T05

    Automation

    Scripts, pipelines and agents that remove repetitive work — and an honest look at what they cost to maintain.

  • T06

    Cybersecurity

    Defensive security for developers: hardening, common mistakes and safe, authorized lab setups.

  • T07

    Developer Tools

    Editors, CLIs and tooling reviewed by actually using them on real work.

  • T08

    SaaS & Product Development

    Build-in-public notes on taking software products from idea to something people can use.

On the bench

Topics AI Techbom is considering for upcoming videos and articles. These are ideas, not published work or promised dates. Finished experiments will appear under Videos and Insights.

  • Idea

    AI agent vs. manual refactor on a real Next.js codebase

    Give the same refactor to an AI coding agent and to a human, then compare time, diff quality and bugs introduced.

    AI Coding

  • Idea

    Building a small AI agent workflow from scratch

    A step-by-step build of a tool-using agent, including the failure modes that only show up once you run it for real.

    AI Agents

  • Idea

    AI coding tools compared on the same task

    One task, several tools, same repo — a transparent comparison with the prompts and results shown on screen.

    AI Coding

  • Idea

    Auditing a Next.js site for SEO and performance

    A practical audit of a real App Router site: metadata, structured data, Core Web Vitals and what actually moved the needle.

    Next.js & TypeScript

  • Idea

    Automating a boring business task end to end

    Pick one repetitive task, automate it, then measure what it really saved after a month of use.

    Automation

  • Idea

    Hardening a contact form: spam, abuse and injection

    Honeypots, rate limits, validation and safe email sending — what each layer stops and what it does not.

    Cybersecurity

  • Idea

    Cloud POS build log

    Build-in-public updates from taking an offline POS application to a cloud-based successor.

    SaaS & Product Development

  • Idea

    One TypeScript trick in under a minute

    A repeatable Shorts format: one concrete TypeScript or Next.js tip, shown working in a real editor.

    Next.js & TypeScript

  • Idea

    An AI agent fail, explained in 60 seconds

    A short clip of an agent going wrong, why it happened and the one-line fix.

    AI Agents

  • Idea

    A developer tool worth keeping

    A single CLI or editor feature demonstrated on real work, with an honest verdict.

    Developer Tools

  • Idea

    A practical checklist before you trust an AI agent with real work

    Written companion to the agent builds: permissions, logging, rollback and human review points.

    AI Agents

  • Idea

    Next.js metadata and structured data, done properly

    A reference for App Router metadata, canonicals, Open Graph and JSON-LD, with the mistakes to avoid.

    Next.js & TypeScript

  • Idea

    Offline-first design lessons from building a POS

    What building a reliable offline POS taught us about sync, conflict handling and failure states.

    Software Engineering