TT5: Partner, Not Replacer
June 4, 2026
This weekly email is my curated selection of interesting and useful topics related to BIM, design, and technology. I aim to provide you with some good information and a few laughs along the way. So, what did I find interesting this week? Read down to find out!
Quick note before we dive in: enrollment is now open for the next cohort of Claude Workflows for Architects, my 8-week program on putting Claude to work on real project tasks. The cohort kicks off June 17th, so if you've been waiting for the next round, this is it. As for this week's links, there's a clear thread running through them: where AI actually fits in our work. From parametric tools to model checking to the question of which parts of your job are automatable, everyone is wrestling with the same question from a different angle.
#1: Twisting Tower Three Times
I built the same twisting tower three different ways to explore a different concept for creating parametric tools. It's one thing to debate Dynamo versus code versus AI in the abstract. It's another to put the same problem through each approach and see what the process actually feels like.
Click here to read the full article at ArchSmarter
#2: Is AI the End of Dynamo?
Jeff from The Revit Kid sits down with Autodesk's John Pierson to dig into what's new in Dynamo for Revit 2027 and whether visual programming has a future now that AI is everywhere. They get into where AI helps, where it falls short, and where Dynamo still wins. A timely companion to this week's #1.
Click here to listen to the episode at The Revit Kid
#3: An Experiment in Model Checking
Jason Kunkel fed Claude a pile of Model Checker checkset XMLs to see if it could build working checks, including a genuinely tricky one comparing door and window fire ratings against their host walls. It worked, but only after he corrected the AI's confidently wrong assumptions about a tool it barely knew. His conclusion matches my own experience: AI is a great partner for niche BIM tools, not a replacer, and it only knows what you teach it.
Click here to read the full article at Design Tech Field Guide
#4: Is Your Job Clean or Messy?
Randy Deutsch argues that the jobs most resistant to AI are the messy ones: ambiguous goals, relational trust, site judgment, and accountability when the data is incomplete. The honest part is his admission that messiness is a slowing force, not a guarantee, and that the better long-term play is being the person who works with AI in messy domains. Worth reading just to inventory which parts of your own job land on which side of the line.
Click here to read the full article at Who We Need To Be
#5: The Rulebook of Your BIM Project
BIM Corner kicks off a new series with a comprehensive guide to the EIR, or Exchange Information Requirement. This document tells your delivery team exactly what data to deliver and when. If you work on ISO 19650 projects or just want tighter information requirements on any project, this is a solid foundation piece. They also touch on automating compliance checks against the requirements, which pairs nicely with link #3.
Click here to read the full article at BIM Corner
That's all from me. Hope you're having a great week.
Michael
