I keep hearing the same worry from friends who aren’t developers: the new AI models sound incredible, but every guide reads like it was written for engineers. So when I stumbled on this breakdown, I got a little excited.
An AI professional took Anthropic’s brand new Fable 5 guide, the one built for coders, and translated it into plain steps for the rest of us. The original poster basically did the hard work of turning developer-speak into a routine you and I can copy today. I was genuinely impressed by how simple the creator made it feel, so I want to walk you through what they shared.
What the expert says Fable 5 actually is
According to this contributor, Fable 5 is the smartest AI on Earth right now, and it behaves nothing like the older Claude you might remember. Instead of blurting out an answer, it works more like a careful senior colleague.
Here’s how the author describes its behavior:
- It plans, asks questions, and pushes back before diving in.
- It checks its own work before handing it back to you.
- It reads the entire chat every single time.
- Cost scales with the conversation: about $0.15 for 1 turn, $6 at 19 turns, $14 at 40 turns.
The creator uses a great analogy: think of Fable as a $1,000/hour senior lawyer, while Sonnet is the $100/hour intern. You bring in the expensive brain for the hard thinking, then hand the routine work to the cheaper helper.
The 5-step routine the author recommends
This is the part I think most people will actually use. The original poster laid out a clean sequence, and each step has a reason behind it.
- Open Claude and select Fable 5 (High). You want the strongest model available before you type a word, because the quality of everything downstream depends on it.
- Give it a goal, not a task, and keep one goal per chat. The expert points out that a single clear goal keeps the model focused and the context clean, instead of it juggling five half-finished jobs.
- Make it interview you BEFORE it starts working. This is the trick I liked most. By forcing questions up front, you catch wrong assumptions early rather than after the work is done.
- Say which mode you’re in. If you’re just thinking out loud, tell it so. If you want it to run the whole thing, tell it that too. Naming the mode stops it from writing a final draft when you only wanted a rough think.
- After 2 turns, switch to Opus 4.8 (High). The author notes you stay in the same chat, the context carries over, and the heavy billing stops. You get the expensive brain for the setup, then a cheaper engine for the follow-through.
The prompt template the creator shares
The person who posted it included a fill-in-the-blank prompt you can paste straight into Claude. Here it is exactly as they wrote it:
“I need [task] for [goal]. I expect [goal] achieved once we hit [specific targets]. Lead with the outcome. Your first sentence answers the question. Details after. Only report work you have evidence for. If it’s not verified, say so. Start by asking me questions to fully understand the context.”
Notice what this prompt is doing. It hands over the goal, defines what \”done\” looks like, demands the answer first, blocks unverified claims, and forces that opening interview from step 3. It’s a lot of good habits packed into a few sentences.
For the two mode signals from step 4, the contributor suggests simple lines like \”Give me your thoughts, don’t write yet\” when you’re brainstorming, and \”Handle it end-to-end\” when you want it to run the full job.
Why the timeline matters
What really caught my attention was how fast this savvy professional says things have moved. In a matter of weeks:
- June 9: Fable 5 launched, included in all paid plans.
- June 12: Pulled offline worldwide, overnight.
- July 1: Brought back, up to 50% of weekly limits.
- July 12: Moved out of subscription plans, so you now pay per use.
The mind behind the post also flags the new pricing: roughly $10 per million input tokens and $50 per million output. That shift from \”included\” to \”pay per use\” is a real change in how you’d budget for it.
Why it matters: the pace here is wild. A model can launch, vanish, return, and change its pricing model in about a month. The expert’s point is that knowing how to steer it well pays off more than ever, because you’re now paying for every turn.
My quick take
What I appreciate about this breakdown is that it’s not about coding at all. It’s a repeatable habit anyone can build: pick the strongest model, hand it a goal, make it ask questions, name your mode, then downshift to save money. Simple, and you can try it in the next five minutes.
The original poster packed even more nuance into the full write-up than I could fit here. Head over to the full LinkedIn post to catch the details in the author’s own words.