What we have learned so far
The tools we use, how we set them up, and what the words actually mean. We wrote it down in case it saves someone else the time.
None of this is settled. It is what is working for us right now, and a good chunk of it will be out of date in a few months. Take whatever is useful.
Part one
What all of this actually means
Nobody agreed on what these words meant before everyone started using them. Here is how we use them.
The whole thing is three layers
Nearly every product you have heard of sits in one of these three layers. We ran them together for a while and it made the whole picture harder to follow than it needed to be.
Where you touch it
Terminal, desktop app, browser, phone, code editor. Doors into the same thing.
The harness, which gives it hands
Claude Code, Cowork, Codex. The layer that decides what the model can see, touch, and change.
The model, which is the brain
Fable 5, Opus 4.8, Sonnet 5, Haiku 4.5, Sol, Terra, Luna. Raw intelligence with no hands and no memory of you.
The part that helped us most: it is the same brain underneath, wearing different gloves. When people argue about which tool is better, they are usually arguing about the harness, not the model.
The vocabulary
Model
The brain. A pattern-matcher trained once on an enormous amount of text, then frozen. It does not learn from your conversation and it does not change until the next release. Everything else here is a wrapper around one of these.
Agent
A model given tools and a loop, so it can act instead of only answer. It reads something, decides what to do, does it, checks the result, and goes again until the work is done.
Harness
The software wrapped around a model that gives it hands: file access, a terminal, memory, and rules about what it may touch. Claude Code and Cowork are two harnesses over the same models.
Context window
Working memory for one session. Everything it can hold at once: your files, the conversation, its own notes. Run out and the older material falls away.
Tokens
The chunks text gets cut into, roughly three quarters of a word each. Context windows are measured in them and billing is counted in them, separately for what goes in and what comes out.
Knowledge cutoff
The date the training data stops. Ask about anything after it and the model is guessing unless it can search or read files. It is why a good agent looks things up instead of answering from memory.
Reasoning
The model working through a problem before it answers, sometimes for a while. It costs tokens and time, and on hard problems it is the biggest lever on quality.
MCP
Model Context Protocol. The standard plug that lets an agent talk to your other software: Slack, Notion, a database, a design tool. It is how an agent acts inside apps it was never specifically built for.
Frontier model
Whatever a lab currently ships as its most capable model. A moving label rather than a product name, and usually the most expensive per token.
Who makes what
Two labs matter most for the work we do. Prices are per million tokens, input cost then output cost.
Anthropic
Makes Claude. This is what we build on.
Claude Fable 5
most capableBuilt for agents that run a long time without losing the thread.
Claude Opus 4.8
our defaultComplex agentic coding and enterprise work. What we build on.
Claude Sonnet 5
The balance of speed and intelligence for everyday work.
Claude Haiku 4.5
fastestNear-frontier answers at the lowest cost, for volume and quick turns.
OpenAI
Makes ChatGPT and Codex. Shipped as three tiers rather than one model.
GPT-5.6 Sol
flagshipThe frontier tier, and what powers the reasoning modes on paid ChatGPT plans.
GPT-5.6 Terra
The balanced everyday model, at half of Sol's price.
GPT-5.6 Luna
fastestThe quickest and cheapest of the three.
Where the tools sit
These are harnesses, not models. Each one can run on more than one brain.
Claude Code
AnthropicThe harness for software. It reads a codebase, writes and runs code, and fixes what breaks. Lives in the terminal, the desktop app, the web, and inside VS Code and JetBrains.
Claude Cowork
AnthropicThe same approach pointed at everything that is not code: research, analysis, documents, multi-step admin. You give it a folder and plain instructions. Desktop, web, and mobile.
Codex
OpenAIOpenAI's coding agent. Searches repositories, runs terminal commands, and debugs. Runs on the GPT-5.6 tiers, with older 5.5, 5.4, and Spark still selectable.
What just changed
The last two months, so the pace is visible.
Anthropic released Claude Fable 5, its most capable widely available model, aimed at agents that run for a long time.
Cowork moved to the cloud and onto web and mobile, so a task keeps running after you close the laptop. Anthropic's own usage data showed most Cowork users were not coding.
Chat and Cowork merged into one Claude. On web and desktop they share a single home screen, and projects and files carry across both. The wall between talking to it and handing it work is gone.
OpenAI shipped the GPT-5.6 family as three tiers instead of one model, and it landed in ChatGPT, Codex, and GitHub Copilot the same day.
Every model and price above was checked against the vendors' own documentation on July 22, 2026. Both labs have been shipping about once a month, so treat this as accurate on that date and not much past it. If something looks off, it probably moved.
Part two
Getting set up
How we connect GitHub, Vercel, and Claude Code. Most of it is making accounts. Plan on about an hour, and it only happens once.
This assumes a Mac and a Claude Pro or Max plan. Copy each command with the button, paste it into Terminal, press Enter.
The three accounts
GitHub and Vercel are enough to start. Supabase comes later, and only if you want it.
GitHub
Where the files live. Version history in the cloud, free, and each project is a repo. Most of the internet runs on it.
Vercel
Hosting. It takes whatever is in GitHub and pushes it live. Push a change and the site updates in about a minute, and every branch gets its own preview link to share.
Supabase
laterThe database, for when a static page is not enough: dashboards, a library of something, a task manager. Worth skipping until you need it. One sentence to Claude spins it up.
In order
- 1
Make a GitHub account
Sign up at github.com with an email you plan to keep. That is enough for now. Once Claude Code is running you can tell it "make all my repos private" and it handles the rest.
- 2
Make a Vercel account with GitHub
At vercel.com, choose "Continue with GitHub." Install the GitHub app when it asks and give it access to all repos. Vercel can now see and deploy anything in GitHub.
- 3
Open Terminal and install Homebrew
Terminal is the app where you type commands. Find it with Spotlight: press Cmd and Space, type "Terminal," press Enter. Paste this in and press Enter. It asks for your Mac password once, and nothing appears on screen while you type it. That is expected.
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"Homebrew is the tool that installs other Mac tools. We reach for it again later for things like ffmpeg, which is what lets Claude edit audio and video.
- 4
Install Claude Code
Still in Terminal, paste this and press Enter. When it finishes, type claude and press Enter to start it. Typing claude --version prints a version number, which confirms it landed.
brew install --cask claude-codeClaude Code runs on a Claude Pro or Max plan, the same login you already have. If you would rather stay out of the terminal, the Claude desktop app does the same work in a window.
- 5
Point it at a folder and talk
Make a folder anywhere (Documents is fine), open Terminal in it, and run claude. From there you describe what you want. It reads the folder, writes the code, runs it, and fixes what breaks.
- 6
Go live when you are ready
In Vercel, open the project's Domains tab and add your domain. Point the main site at something like yoursite.com. Later you can add a private dashboard at a subdomain like admin.yoursite.com to run projects from, which is where Supabase comes in. Nothing is public until you point the domain.
How we work
Live in the agent, not a browser
We open a browser to preview something and not much else. The rest happens in the conversation.
Point it at a repo, not a chat project
"Open my [name] repo with all the context" saves re-explaining yourself at the start of every session.
Give it house rules
A CLAUDE.md file holds how you write and what you never want it to do. Saying "add this to memory" works too, and it sticks.
Keep one repo as the source of truth
Ask it to interview you, something like "ask me 100 questions about me and my work," and it will organize what it learns. That becomes the thing you point everything else at.
Start from something that exists
Reusing a past build beats starting cold, and it saves teaching the same preferences twice.
Read before you accept
It is easy to rubber-stamp the seventh prompt in a row. We have done it and regretted it. A quick skim of what it is about to do is usually enough.
When you want more
Getting connected is the first half. Building your own back office on top of it is where it gets interesting.
Skills. Packaged workflows for jobs you repeat.
Loop. Have it wake up on an interval and keep improving something until it is done.
Artifacts. Turn words into a clean page, like a proposal with signatures built in.
Usage. Type /usage to see what you are actually spending.
Part three
What we use
The tools we run our businesses on, with our referral codes in one place.
Everyday tools
These carry our referral code, so everybody gets something.
Granola
Takes the notes in every meeting we are in so we can actually be in the meeting. The summaries are good enough to send to clients.
Wispr Flow
Voice dictation that keeps up with how people actually talk. Most of our emails and messages are spoken first now.
Mercury
Business banking that feels like software instead of a bank lobby. Every company we run banks here.
Raycast
The command bar for your Mac. Launcher, clipboard history, window management, and quick AI, all from the keyboard.
Build agents
The ones doing the heavy lifting. No codes here, only what works.
Claude Code
Agentic coding tool that lives in your terminal. Describe what you want, it reads the codebase, writes the code, runs it, and iterates. Does most of the heavy lifting around here.
Codex
OpenAI's coding agent, same job as Claude Code. Running both means work happens in parallel and you get a second model's take on tricky problems.
Claude Design
A design agent with a visual canvas. You iterate on layout and look through chat instead of hand-placing everything, then pull the result into the build.
The stack
What every site and app we ship runs on.
Next.js
The framework. React for the component model, Next.js for routing, server rendering, and the build pipeline.
shadcn/ui
UI components built on Radix and Tailwind that you copy into your repo, so you own the code and can change anything.
Vercel
Hosting and deploys. Push to GitHub and it is live in a minute, with a preview URL for every branch.
Worth reading
Ideas we keep coming back to.
The Only Question Left
The cost to think and the cost to build are both heading to zero, which leaves one question: what is worth building. Not ours, but it makes the case better than we would have.
The short version: we describe what we want and the agents do the heavy lifting. Everything lands in GitHub, every push deploys through Vercel, and Supabase holds whatever needs a database. Around the edges we talk more than we type. That is most of the trick.