Part 1 - “The Future of BIM is NOT BIM”
“The Future of BIM is NOT BIM” presentation by EvolveLAB CEO Bill Allen showed the impact of reducing friction throughout sectors of AEC and encouraging the AEC industry overall that the time to bridge the data gap between modeling,design, and construction is now. Allen’s examples of “reducing friction” spanned from existing developments like the Dynamo player, which makes it simpler for users to run Dynamo scripts, and businesses using LiDAR to create as built models of projects. He emphasized user experience and making workers more efficient by removing some of the tedium from their roles and thus increasing the productivity of existing workers and existing tools. I would definitely be interested in tools that introduce automation for labeling things in Revit and making sheets.
For the threat to human workers, I wasn’t completely convinced. I did note that, in making his argument that technology using AI, generative design, or robotics isn’t going to take your job, he showed a statistic that we’re short on labor in the construction industry. Is the same true for designers? That being said, I accept that the world is changing and has changed before. I just worry about what that’s going to look like.
He ends his presentation with a quote from Shajay Bhooshan: “Design cannot be simply a matter of intuition when you’re solving complex problems”. I once watched a video of an architect who was talking about how the rise of modeling tools like Revit fundamentally changed how we design for the worse. While I believe in optimization and believe there is creativity that can still be incorporated once you have the best building forms or configurations from dynamo, I do wonder if we lose something creatively. (Seeing as Bhoosan comes from Zaha Hadid Architects, I’m curious to look more into how they are incorporating big data and AI.
Part 2
I was trying to adapt some logic from my work in Module 6 to incorporate it in Module 7, and the process was way more painful than I anticipated. One specific area of friction was debugging the differences in logic for a form created in Revit from a form created in Dynamo. Even file organization and understanding why Dynamo’s having trouble locating a file were big issues here. Allen’s note in the presentation about the difference in digestibility for data in spreadsheet form vs graphs. Though I’ve become accustomed to reading and understanding spreadsheets, when I hand that information on to a client or supervisor, there are definitely better forms for it to be in and it will reduce friction in data communication.
I think AI tools will be very helpful in allowing people to adapt their existing work or, with credit and permission, better utilize other people’s existing code and tools. These are cases where the intent and logic have already been sorted out, but you may have implementation difficulties where AI can assist. Of course, the user still needs to understand enough to know that the code is still doing what they intend and that the solutions being output make sense. I try to use AI pretty sparingly as I work on developing skills in school, but MATLAB has an AI feature for debugging your existing code that I’ve definitely found useful. Incorporating something similar in Dynamo would be interesting. Of course, there’s already error messages, but I did find myself having to look online for what they really meant.
Part 3- Find three tools, companies, projects, research or open-source efforts working at the intersection of AI/ML and the built environment.
Spectral Labs is a company that has made a generative model for CAD. The model is called SGS-1. It’s a plug-in for Fusion360 where you can type in prompts, and it produces an element based on the prompt in parametric geometry in CAD. It can also be used to convert a sketch or engineering drawing into CAD geometry. This isn’t directly usable for what I was doing in class, but I imagine this saves a huge amount of time for people in the design process.
https://www.sciencedirect.com/science/article/pii/S0360132325005359
This research paper looked at incorporating daylight performance into the evaluation of AI-generated floor plans. Their methods in the study improved the efficiency and simulation time of the generation of floor plans over all. This could be a helpful tool for designers who want to quickly determine viable floor plans with good daylight penetration.
Autodesk Assistant is an integrated agentic AI chatbot in Autodesk products. From my understanding, it is currently only being used in Revit to assist search in the Help function of the application. It seems the goal as this product develops is that it can streamline the connection and transfer of models between Autodesk products. The marketing materials for the assistant say their goal is to allow the user to stay in a creative mindset and not get bogged down by the tool.