I lost count of how many times I've typed a lazy prompt, hit enter, and quietly hoped the AI would read my mind. It almost never works.
So when I found a breakdown of Anthropic's brand-new 31-page prompting guide, I stopped scrolling immediately.
Someone took that dense document and squeezed it into 10 rules you can use today, tuned for how Opus 4.8 actually behaves. The breakdown is genuinely practical, so let me walk you through it.
In partnership with
Talk to your AI tools the way you'd talk to a colleague.
You don't send a colleague a three-word brief. You explain the context, the constraints, what you've already tried. But typing all that into ChatGPT takes forever — so you don't.
Wispr Flow lets you speak your prompts instead. Talk through your thinking naturally and get clean, paste-ready text. No filler words. No cleanup. Just detailed prompts that actually get you useful answers on the first try.
Millions of users worldwide. Works system-wide on Mac, Windows, and iPhone.
The one change that rewrites everything
Here's the shift the whole list hangs on. Older Claude versions guessed what you meant. The new 4.8 does exactly what you type.
That one change rewrites how you should prompt. Less mind-reading, more instructions. Here's the full list.
The 10 rules, broken down
Name every output. Instead of "review this contract" and praying, spell it out: "Review this contract. Flag risks per clause. Rate severity 1 to 5. Return a table." You get structure because you asked for structure.
Cap the length yourself. "Summarize this" on a 40-page report lets the model size the answer to the input. Box it in: "5 bullets. Each under 15 words. Start each with an action verb."
Flip your negatives into positives. "Don't use jargon, don't be salesy" doesn't stick. State what you DO want: "Write in plain English a 16-year-old could read aloud."
Make every verb ship something. "Can you help me with the email?" is too soft. Turn it into a chain: "Open Gmail. Find [contact]. Write the send-ready reply. Under 90 words. Tone: confident, casual."
Force the web search. Opus 4.8 calls fewer tools than 4.6, so you have to push it: "Use web search aggressively. Verify every claim with at least 2 sources."
Feed it your voice. 4.8 is direct, with almost zero emojis. Miss the warmer tone? Paste 2 or 3 sentences in the voice you want, then tell Claude to match the rhythm.
Ask it to go further. Ask for "a landing page" and you get the bare minimum. Steal this line straight from Anthropic's doc: "Go beyond the basics." Drop it on every creative task.
Turn on thinking. Claude 4.8 doesn't reason by default. Add this to the end of your prompt: "Think before answering (maximum reasoning)." A free upgrade, every single time.
Build a skill from repeat prompts. Rewriting the same prompt 14 times a week? Stop. A skill is a command with the instructions pre-built. Write the same prompt twice, make it a skill.
Spell out everything. Old Claude guessed. New Claude does exactly what you typed, so define your output, order, length, tone, and format. If you don't say it, you don't get it.
How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads
The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.
*Ad
Why this list actually matters
Most of these rules trace back to one behavior change in 4.8: it stopped filling in the gaps for you.
That's not a downside. It means you finally get predictable output, as long as you're specific. The lazier your prompt, the more you're gambling. The clearer your prompt, the more the model just does what you asked.
Specificity is the whole game now.
3 things to try in the next five minutes
If you only steal three things from this list, make it these.
Add "Think before answering (maximum reasoning)" to your next complex prompt and watch the quality jump.
Replace one negative instruction today with a positive version of what you actually want.
Define the output format upfront, whether that's a table, five bullets, or a word count.
Try them on something boring, like cleaning up an email or summarizing a doc. You'll feel the difference fast.
The habit that's quietly costing you
We spent years training ourselves to write loose prompts and let the AI guess. With Claude 4.8, that habit costs you quality.
Tightening up how you ask is the single highest-leverage skill right now. Not a bigger model, not a secret phrase. Just saying what you want, clearly, the first time.
Payroll errors cost more than you think
While many businesses are solving problems at lightspeed, their payroll systems seem to stay stuck in the past. Deel's free Payroll Toolkit shows you what's actually changing in payroll this year, which problems hit first, and how to fix them before they cost you. Because new compliance rules, AI automation, and multi-country remote teams are all colliding at once.
Check out the free Deel Payroll Toolkit today and get a step-by-step roadmap to modernize operations, reduce manual work, and build a payroll strategy that scales with confidence.
*Ad
Go grab the full list
The original breakdown lays out all 10 rules with before, fix, and a copy-ready example for each. That's rare. Most prompt advice is theory with no receipts.
Save it somewhere you'll actually find it again, then run one rule on your next prompt. The boring tasks are where you'll feel it first.
Quick poll
Which rule are you stealing first?
