The 'usage limit reached' banner always arrives at the worst moment: mid-task, mid-flow, mid-good-idea. I have hit that wall more times than I want to admit. One creator on LinkedIn got tired of their team slamming into Claude's weekly cap, so they wrote down every token-saving habit they could find.

They called it 22 rules; the real count is 24. The original poster says up front they are not paid by Anthropic, and this is just what worked after weeks of hitting the ceiling. The rules split into clean buckets: settings you pick before typing, how you run the chat, what you feed the model, and the small daily habits that pile up unnoticed.

Here is why I find this worth your time: most token burn has nothing to do with your prompts. It hides in the background, in files, tools and instructions you forgot were loaded. Fix those, and the same plan stretches across the whole week instead of dying on Thursday.

In partnership with

The agentic era needs a different CRM. That’s Attio.

Teams like Parallel, Turbopuffer, and Wordsmith are already setting the pace on Attio. Get an always-on revenue engine, with agents and workflows that build pipeline, chase every buying signal, and move deals forward with your team. Whether you're working in your browser, inbox, or favorite agent, connect to your customer data in real-time through Attio's web app, MCP, API, and SDK.

1. Pick the settings before you type

The fastest wins come before you type a word. The author's first move: default to Medium effort, never Max. Higher effort means more thinking tokens on every reply, and most everyday work runs fine without the deep reasoning.

Then match the horsepower to the job. Low effort for grammar, high for one big, complex prompt. The job sets the dial.

Two more settings rules stuck with me. Pick one strong default model and stay in it instead of grabbing the heaviest option every time. And never switch models mid-chat: the new one reprocesses the whole thread from scratch, and that burns tokens you will miss.

2. Short chats beat long ones

Here is the mechanism that reframes everything: every message you send makes Claude reread the entire thread. Long conversations get more expensive with each reply. The cost compounds.

So the creator's advice is blunt: never reopen an old chat. That buried history gets reprocessed every time you add a line. Ask for a 10-line summary, paste it into a fresh chat, and keep the context without the dead weight.

One topic per chat, too. Mixing subjects bloats the window with stuff unrelated to your real question. And resist correcting the model: editing your original prompt replaces the bad round instead of stacking another one on top.

Two habits I am stealing right away. Put every ask in one message, since one full request costs less than five nervous follow-ups. And open big tasks with 'ask me questions before you start,' because a few questions up front beat three rounds of redoing work the model guessed wrong.

Knowing these habits and actually running them are two different things, and the second only happens with reps. The claude track in 3 Minute AI gives you those reps in three-minute lessons, each ending with a task you run in a built-in chat lab. Daily lessons and two full courses are free.

Turn today's calls into tomorrow's proposal

Wispr Flow Notetaker brings your meetings into Claude or ChatGPT through a built-in connector. Ask for a proposal draft or a deal recap, and your AI works from an accurate record of what was said.

*Ad

3. The hidden drain is everything you feed it

This section was the real eye-opener for me. A huge share of token burn has nothing to do with prompts and everything to do with the files, tools and instructions idling in the background. You pay rent on all of it.

End quick questions with No web search. No files. That one line stops the detour.

Crop every screenshot and turn PDFs into text, because full images and documents cost far more than the few lines you really care about.

Then there is the connector tax. Every app you plug in adds its tool definitions to your context, even the ones you never call. Switch them on only when you need them, and keep custom instructions under 2,000 words, since those load into every conversation you start.

Reused docs belong in a Project, uploaded once, not pasted into every new chat. The expert even points to two specific sentences from Anthropic worth dropping into your global instructions. That detail tells me this list came from real pain, not theory.

4. The Sunday max day flip

Now the rule the author loves most, and yes, so do I. The claim: limits reset every Sunday at 5pm, so they turned Sunday into their 'Max day.' They spend every leftover token on the biggest, most ambitious tasks before the counter zeroes out.

I love this because it flips the whole mindset. Unused tokens vanish at the reset anyway. Hoarding them is pure waste.

You stay lean Monday through Saturday, then go big on Sunday with the refactor or the long draft you kept dodging. One caution from me: check when your own account resets before you build a ritual around it. The timing is not universal.

A few daily habits round it out. Plan a deck in plain text first and ask for the file only once, since generating files is expensive. Pause automations you never read, split your work into three sessions so your usage window can recover, and in Claude Code, run one task then clear the context.

If three rules are all you try: default to Medium effort, start fresh chats with a 10-line summary, and kill the connectors you are not using. Those three alone plug most of the hidden leak.

100+ ChatGPT Prompts to Revolutionize Your Day

Supercharge your productivity with HubSpot's comprehensive guide to Chat GPT. This free resource is your fast track to AI mastery:

  • Industry-Specific Use Cases: 15+ real-world applications across various sectors

  • Productivity Guide: 21 best practices to 10x your efficiency with AI

  • Prompt Powerhouse: 100+ ready-to-use prompts for immediate implementation

  • Challenge Buster: Overcome common AI hurdles with expert strategies

Plus, in-depth sections on email composition, content creation, customer support, and data analysis.

*Ad

Ten minutes in your settings tonight

Open your settings today, drop your default effort to Medium, and turn off every connector you have not touched in a month.

When you want the rest, here is the full 24-rule breakdown.

Worth 10 minutes if you keep running dry mid-week and want to stretch the same plan all the way to Sunday.

Five reads to squeeze more out of your plan

Want to go past the 24 rules? Start with Anthropic's own advice, then see how other heavy users keep their week alive.

Hit reply and tell us why.