Human in the Loop – TGN rigs in models for human engagement, guidance, and limits on AI-generated architecture and engineering and models

I wake up early this morning and see again more amazing posts (on LinkedIn) about Mid Journey V4. As that moves more and more into the design of architecture and engineering models, AFR/TGN rigsand rig placement could even be automated by mid journey or anyone’s TGN implementationwould be a very logical and productive form of engagement for humans in the loop, humans who need to tune up the mid journey results and improve them.

Mid journey in fact could be tuned to interpret this human engagement at the TGN rigs and run its processes again taking these human guidances and limits into account.

This of course could be another way to envision software development with AFR/TGN features built into any softwares that handle visual spatial modeled environments.

Typical forms of human engagement with models are targeted at humans in the loop formulating and answering these kinds of questions:

  1. How do you know when, or if, when creating your model, that the model’s “done”? That it’s good enough?
  2. How well does your team understand its own models, or models it receives from others? Do they understand superficially or in depth?
    • If they understand thoroughly, how did they develop that level of understanding? What kind of process do they go through in order to reach that level of understanding? (the answer always involves some form of systematic long term engagement with the models)
  3. Are your models reliable, well built, complete, and correct more so in SOME areas than in others? WHERE are those areas? How can you, or anyone else, know where those areas are in the model?
  4. How, using what means, can you AFFIRM to others that HERE (and here, and here, and there, and there…) that those statements are correct, that the model meets standard of care (reliable, well built, complete enough, correct) HERE.
    • Not to mention the broader question of whether or not the model makes sense at a schematic level, at the level of the building program.

Here’s a detailed discussion and proposal for AFR/TGN features development in modeling softwares (including any/every model-handling software):

Human in the Loop © Hayward Wilkirson
%d bloggers like this: