01. Future of Documentation

Disclaimer: There is no doubt about it that LLM's have been leveraged to write down the Deternet project structure even though it ground work was already created before LLM's were available to the general public. As with any sector LLM's have deeply changed the documentation landscape in many different ways. Today and even more in the future documentation in human readable form for any project will be written by AI. The job of the human expert is then to check if what the AI has written is actually correct. 

Hence, you will not get rid of human technical writers but the job changes from being a writer to being a controller if everything is correct and adjust if something is wrong. You could compare it to a pilot that always flies on automatic pilot but needs to constantly check if everything goes according to plan and adjust if something goes wrong.

To better illustrate this we can look at how LLM's can now write documentation for code that has been created by another AI with a human in the loop. The workflow goes like this in the sense of spec driven development with documentation included.

 

 

Diagrams, Graphs, Work Flows, Etcetera

Instead of you manually taking data and making diagrams, graphs, work flows, charts and other forms of representing data you can let your 

Analysis of the data

All the data created can be used together to look for correlations and you can easily do comparisons with different years, percentages, and take extra data from the internet as well to compare with multiple countries for instance or how would this data likely change if we increased variable x or included variable y.

Documentation for Network Engineers

Yes, all documentation is checked by humans before it is published on any Deternet society or hub. But people in the field will more and more use AI agents that can 

Monitoring Systems

People in the field will have the skills to lay the cables, and to splice the fiber, connect it to the right ports, but more and more can be standardized like OPNSense routers, switches, wap's and more. All this knowledge needs to be collected in a central place and updated constantly. 

How is information retrieved

Most often engineers will not move to any of the documentation pages if the documentation is already included in a large model or small models can be created completely focused on laws and regulations inside a country or region. Basically, the AI is reading the docs and people have conversations with documentation instead of manually searching through it. Even though, myself I catch myself manually search most of the time anyways.