Setup
Setup
The hardware, software, and tools I use every day — from the M4 MacBook Air to the AI agents that ship with me.
Hardware
Devices I use every day
- MacBook Air M4
Daily driver — M4, 24GB unified memory, 10-core GPU. Light enough for travel, strong enough for Docker, TypeScript monorepos, and local models.
- iPhone 16 Pro
Mobile hub — messages, camera, quick reviews, and on-the-go AI chats. Action Button wired to useful shortcuts.
- Mac Mini M4
Always-on home server — Ollama, local LLMs, vector DBs, and long-running agent jobs without taxing the laptop.
- Windows PC
Secondary desktop for Windows-only apps, dual-boot experiments, and occasional GPU-heavy work.
- Apple Watch
Health, focus timers, and glanceable notifications without pulling out the phone every five minutes.
- AirPods Pro
ANC for deep work, spatial audio for reviews, and seamless handoff between Mac and iPhone.
Operating Systems
Primary and secondary environments
Editor & IDE
Where code gets written
- VS Code
Main editor — Prettier, ESLint, Tailwind IntelliSense, Astro, GitHub Copilot, and a tight keybinding set.
- Claude Code
Agentic coding in the terminal — multi-file refactors, PR flows, and long-running implementation loops.
- Cursor
AI-native IDE when I want inline edits, multi-file composer, and a VS Code-compatible UX.
Terminal & CLI
Shell, package managers, and local tooling
- iTerm2 + Oh My Zsh
Terminal home base — Powerlevel10k, autosuggestions, split panes, and years of aliases.
- Homebrew
macOS package manager — the first install on every new machine.
- Bun
Default JS runtime and package manager for Next.js / Astro work — fast installs, fast scripts.
- uv
Python package and project manager — pip-compatible speed without the pain.
- Docker
Local Postgres, services, and reproducible environments without polluting the host.
- Git + GitHub CLI
Version control and PR workflows from the terminal — `gh` for reviews, checks, and stacks.
AI Tools
Models and agents in daily workflow
- Claude
Primary reasoning model — architecture, code review, long-form writing, and agent orchestration.
- ChatGPT / OpenAI
OpenAI models and APIs for product features, image gen, and fallback reasoning.
- Grok
xAI’s models for real-time context, image work, and agentic build loops in Grok Build.
- Gemini
Google’s multimodal models — large context, image generation, and research passes.
- GitHub Copilot
Inline completions in VS Code — boilerplate, tests, and small refactors without leaving the file.
- Ollama
Local model runtime on Mac Mini — private inference, RAG experiments, and offline demos.
Browser
Web browsing and research
Productivity & Design
Notes, design, and collaboration
- Obsidian
Long-term knowledge base — markdown notes, links, and second-brain structure.
- Notion
Shared docs, project hubs, and lightweight databases for teams and clients.
- Apple Notes
Frictionless capture on iPhone and Mac — shopping lists, meeting scraps, quick links.
- Raycast
Launcher, clipboard history, window management, and custom scripts — Spotlight on steroids.
- Figma
Design systems, UI reviews, and design-to-code handoffs for product work.
- Canva
Fast social graphics, decks, and branded visuals without opening a full design suite.
- GitHub
Source of truth for code, PRs, Actions, and open-source repos.
- Google Drive
Shared files, Docs/Sheets collab, and client deliverables that need universal access.
- Slack
Team chat, client channels, and integrations for delivery workflows.
- Jira
Issue tracking and sprint boards for larger delivery engagements.
- 1Password
Password manager, SSH keys, and secrets shared safely across devices.
Deploy & Infrastructure
Hosting, DBs, and cloud
- Vercel
Default deploy target for Next.js and Astro — previews per PR, edge network, Analytics.
- Cloudflare
DNS, CDN, Workers, and non-Next stacks that need global edge without a full Vercel setup.
- Supabase
Postgres, Auth, Storage, and Realtime for live apps — production DB without ops theater.
- Upstash
Serverless Redis and rate limiting — caching and queues without managing Redis hosts.
- n8n
Workflow automation — client intake, mailbox processing, Jira/Monday bridges, AI-assisted ops.
- AWS & Azure
Enterprise cloud when clients require it — secrets, storage, OpenAI on Azure, Node middleware.