Skip to content

What it gets wrong

The honest limits. Read this before you bet a quarter of your content budget on the approach.

5 min read

Level 0

Curious

0 XP
150 XP to ApprenticeNo streak yet0 badges0% done

Every course like this has a lesson that oversells. This is not that lesson. Here is what actually goes wrong, so you can plan around it instead of discovering it in month three.

It will state a wrong number with total confidence

Ask what a tool costs and you will get a price. The price will be formatted correctly, sound reasonable, and be from 2023.

There is no tell. It does not hedge on the invented ones. The only defense is a rule: never state a price or a statistic that was not fetched in this session, and write down the date you fetched it.

It cannot see its own repetition

Inside one article it is fine. Across four, it has no idea. It is not being lazy, it simply cannot read the other three while writing this one. You have to make that comparison a separate step.

It follows the last thing you said

Tell it to be more concise and the next section gets clipped. Tell it to add detail and the section after that bloats. It over-corrects, every time, in the direction of your most recent instruction.

This is why written rules beat live steering. A file says the same thing on message forty as it did on message one.

It trusts its own tools too much

This one is the most expensive, and it gets a full lesson in Module 3. The short version: when a check comes back with a strange result, the first suspect is the check, not the world.

What went wrong2026-07-21

Fifty-four dead links that were all alive

A link check ran over 54 URLs. Every single one came back as unreachable, including several confirmed working minutes earlier in the same session. It looked like the whole client site had gone down. It had not. The list of URLs had been written by a script on Windows, which quietly turned every line ending into two characters instead of one. Every URL carried an invisible extra character, so every request went to an address that does not exist.

Cost: Most of an afternoon, and it nearly triggered a mass 'fix' of links that were fine.

The rule that came out of it. A uniform failure is almost always your tooling. A real outage is selective. When 100% of anything fails, suspect the harness before you suspect the world.

So when is this the wrong tool?

SituationVerdict
One article a month, one writerNot worth the setup. The system pays off across batches.
Writing that depends on interviews you didGreat fit, but you do the interviews. It cannot source original quotes.
Anything where being wrong is expensiveOnly with a verification step you actually run. Medical, legal, financial.
Four to forty pieces on one topic clusterThis is the sweet spot. Consistency is the hard part and this is what fixes it.
You want to stop caring about qualityWrong tool. It raises the floor, it does not remove the need for judgment.

Worth saying plainly

This setup makes a competent writer much faster and much more consistent. It does not turn someone who cannot tell good from bad into someone who can. Every check in this course still ends with a person deciding.

Finished reading?

Mark it done to bank the XP and keep your streak alive. Any challenges on this page score separately, so you can come back for a perfect run later.