Reading Time: 9 minutes

Hey Prompt Lover,

I wasn't expecting that.

After the last newsletter went out, my inbox filled up fast.

People wanting the full Cowork setup. People asking about Claude Code. A bunch of you specifically asking about the phone to laptop thing — how does that actually work, how do I set it up, walk me through it.

And honestly? That told me everything I needed to know about what to do next.

So here's what's happening.

I'm turning this into a proper series. Each issue is going deep on one thing. Real setup instructions. What actually works. What breaks. No surface level stuff.

First up is going to be Cowork — the full setup, the three files, the folder structure, how to stop getting generic output. Then Claude Code for the developers and the people who want to go further. Then the phone to laptop bridge, how to set up Dispatch, how to message your desktop from anywhere and come back to finished work.

Each one is its own guide. Each one is worth your time.

But before all of that, I want to say something I probably should have said more clearly in the last issue.

61% of marketers say this is the biggest marketing shift in decades.

Get the data and trends shaping growth in 2026 with this groundbreaking state of marketing report.

Inside you’ll discover:

Download your 2026 state of marketing report today.

Get Your Report

I use Claude every single day.

Not occasionally. Not when I feel like it. Every. Single. Day.

And in my honest opinion — the kind of opinion you only get from someone who's actually sat with multiple tools for months and stress tested all of them — Claude is the best AI available right now.

I know that's a conversation starter. Some of you will agree, some of you will push back, and that's fine.

But when I look at what Anthropic has shipped in ten weeks, when I look at how Sonnet handles real work, when I look at what Cowork is doing for the people who've properly set it up — I don't know what else to call it.

The gap is real. And it's growing.

Which is exactly why what I'm about to share matters.

Because here's the problem.

A lot of you are going to start taking Claude more seriously after that last issue.

You're going to open it up, start running it hard, and then hit a wall.

The usage limit wall.

It happens to everyone. You're mid-session, things are going well, and suddenly you're being told to slow down or wait. And it's the most frustrating thing because you were just getting somewhere.

So while you're waiting for the deep dive guides, I want to give you something useful right now.

Eight things that will stretch your Claude usage further than you thought possible. Things I wish someone had told me earlier.

Let's go through them.

Edit your prompt. Don't send a follow-up.

This is the one that changes everything once you know it.

When Claude's answer misses the mark, most people type another message trying to fix it. Don't do that. Click the edit icon on your original message, fix the prompt, and regenerate.

The old exchange gets replaced instead of added to the conversation history.

Over ten rounds of back and forth, this one habit alone cuts your token use by 80 to 90 percent. That's not a small number. That's the difference between hitting your limit at noon or making it through the whole day.

Fix the prompt. Don't stack the chat.

Start fresh every 15 to 20 messages.

Here's something nobody talks about.

Claude re-reads the entire conversation history every single turn.

Your first message costs around 200 tokens. By message 30, a simple follow-up question can cost 50,000 tokens or more. The chat gets heavier every time you add to it.

When you feel a conversation getting long, ask Claude to summarize what's been decided and what the key points are. Copy that summary. Open a new chat. Paste it in and keep going.

Fresh context. Same momentum. A fraction of the cost.

Long chats are expensive chats.

Batch your questions into one message.

Instead of sending three separate messages, combine them.

"Summarize this article, pull out the main points as bullets, and suggest a headline."

That's one turn instead of three. One context load instead of three. And the answers are often better because Claude sees the full picture of what you need before it starts.

Three questions. One message. Always.

Upload your recurring files to Projects.

If you're uploading the same PDF, the same brief, or the same guide across multiple chats, Claude is counting those tokens every single time.

Projects in the sidebar cache your files so they don't cost again on each new conversation.

If you work with long documents regularly, this one change alone saves you a significant chunk of your daily budget.

Upload once. Stop paying every time.

Set up Memory and Custom Instructions.

Every conversation you start without any context burns three to five setup messages just re-explaining who you are and how you work.

Go to Settings, then Memory and User Preferences, and store it once. Your role, your tone, what you're working on, how you like responses formatted. Claude carries it automatically into every chat after that.

You stop explaining yourself from scratch every single session.

Set it once. It runs forever.

Turn off features you're not using.

Web search, Research mode, and connectors all add tokens to every response — even when you're not using them for that particular task.

When you're working with your own content or just writing, toggle off Search and tools.

Extended Thinking is the same. Leave it off by default and only switch it on when your first attempt wasn't good enough and you need Claude to work harder.

If you didn't turn it on intentionally, it should be off.

Use Haiku for simple tasks all day.

This is the single highest impact decision most people aren't making.

Haiku 4.5 handles grammar checks, quick answers, brainstorming, formatting, and translations at a fraction of the cost of Sonnet or Opus.

Using Haiku all day for simple work frees up 50 to 70 percent of your budget for the tasks that actually need the bigger models.

Think of it like this.

Quick answers, brainstorms, formatting, grammar — that's Haiku. Very low cost.

Content writing, analysis, coding, drafts — that's Sonnet. Medium cost.

Deep research, hard logic, long document review — that's Opus. High cost.

Haiku for drafts. Sonnet for real work. Opus for the hard stuff.

Most people run Sonnet on everything. That's like driving a truck to buy milk.

Spread your work across the day.

Claude runs on a rolling five-hour window that resets continuously.

If you burn through your limit in one morning session you're done until the window rolls over. But if you split your work into two or three sessions across the day instead of one burst, you can effectively get 150 to 200-plus messages a day on a Pro plan instead of 45.

Don't sprint. Pace yourself.

The limit isn't the ceiling. How you use it is.

That's eight things you can apply today before the guides even drop.

And the guides are coming.

Cowork setup is first. Full walkthrough, the files, the folder structure, the exact prompts that work. Then Claude Code. Then the phone to laptop bridge.

If there's something specific you want covered in each one, reply and tell me now. The best questions shape what goes in.

See you in the next one.

— Prompt Guy