In partnership with
👋 Welcome to AI Central,
I was working with AI on Sunday and ran into this annoying problem. Every time I asked it to fix something, it would mess up something else. And then I’d have to ask it to fix that too. Before I knew it, I was going in circles.
So I looked into how to avoid this and thought I’d share it with you in today’s issue. Hopefully, you can save yourself the same headache!
We’ll look at how to get AI to understand what you want from the start, and when it’s better to stop making changes.
Big hugs ❤️
Kris & The AI Central Team
P.S. Don’t forget to download our bonus guide on 20 ChatGPT mistakes to avoid - find it below!
🔗 TOGETHER WITH THE CODE
100+ Claude Code hacks to ship code 10X faster
Top engineers at Anthropic and OpenAI say AI now writes 100% of their code.
If you're not using AI, you're spending 40 hours doing what they do in 4.
These 100+ Claude Code hacks fix that and help you ship 10x faster.
Sign up for The Code and get:
100+ Claude Code hacks used by top engineers — free
The Code newsletter — learn the latest AI tools, tips, and skills to code faster with AI in 5 minutes a day
You're seeing this ad because you're on the free plan - upgrade to remove ads & unlock 1200 AI Tutorials. Ads keep AI Central free for everyone
📚 The guide you need next is probably already written
Whatever you hand to AI next (a contract review, a sales email, a weekly report), someone on our team has likely already spent the hours testing how to do it well, and written it up.
The AI Central Library puts those guides in one place. Search by tool, task or author, save the ones you'll reuse to your own playlists, and add notes as you go so the useful part is there next time.
Your first month is $4.99.
🔗 BROUGHT TO YOU BY TLDR AI
Why 1.1M+ readers open this AI newsletter every day
From ML engineers to founders, 1.1M+ readers use TLDR AI to spot new tools and research before they go mainstream. One free email, every morning.
You're seeing this ad because you're on the free plan - upgrade to remove ads & unlock 1200 AI Tutorials. Ads keep AI Central free for everyone
📓 Judge AI output like an editor
Inside this guide you will learn:
Section | What you can do by the end |
|---|---|
22.1 True, sound and good are three different questions | Run the right check so a draft can't pass two of them and fail the third |
22.2 Write down what good looks like before you read a word of it | Write a 4-line bar before anything exists to influence you |
22.3 Change one thing, and go back rather than forward | Run a loop that converges, and never lose the version you liked |
22.4 Stop when describing the change costs more than making it | Pick which of the 4 reasons to stop applies - only one means it's finished |
You'll need: one draft the assistant produced that you weren't happy with but couldn't say precisely why. That draft is the working material for the whole chapter.
📒 20 ChatGPT mistakes to avoid
This guide breaks down 20 common mistakes that can lead to generic, inaccurate or unhelpful AI responses. From writing vague prompts and cramming too many requests into one to ignoring custom instructions, skipping examples and trusting every answer, it shows you what to fix.
Use it as a practical checklist to improve your prompts, get better results from ChatGPT, and build a more effective AI workflow.
