AI Guide
How to assign tasks, check results and avoid unnecessary work.
Choose a topic to explore the guide.
Guides and case studies
9 guides
Why AI invents even a simple path through settings
A usage-counter example, an explanation of hallucinations and an interview with Professor Aleksander Mądry from 1:11:52.
A good brief. A verified result.
How to describe a task, provide the necessary context and check the AI's work.
One chat or separate projects?
When to stay in a conversation and when to start a separate project to preserve continuity.
Same instruction. Different results.
Complete task histories, responses and nine panels. Findings and limits of the comparison.
How to organise work with AI
Rules, current task status and a project document: a simple division of responsibilities.
When AI creates its own bottleneck
The registry grew to 336.7 MB. What we simplified, what the measurements showed and how to prevent a repeat.
Why AI opened empty CC and BCC fields
How checking recipients turned into unnecessary clicking and how the procedure was improved.
When an attachment fails, check the message
An unreadable attachment is not the end of the source check. The sequence, three screenshots and a practical lesson.
Hidden rows or a scrolled spreadsheet?
Off-screen rows were mistaken for hidden rows. A mistaken report, two user corrections and the fix.
Conversations with AI
Translated conversations · models · response times
An interesting conversation between nobody and AI nobody about creativity
All the world’s knowledge, but where is the idea? A human questions AI about creativity
Pressure for results and AI reliability
When pressure encourages apparent completion and how to demand verified results.
Is AI creative? Part II
A new pattern is not necessarily creativity
