More than 8 years ago. I made some predictions on this blog on the possible direction of technology within design & construction.
To put in in context, at that time: Australia didn’t have a national broadband network. iPhones still had a 3.5mm headphone jack. WFH wasn’t a thing. We were laughing at, not despising Donald Trump. There was a bit of hype about ‘BIM’ but nothing about Artificial Intelligence.
Of course, I was wildly optimistic on some things, but didn’t do too badly all things considered.
Cloud 
Delivering a major construction project without applications such as BIM360/Autodesk Construction Cloud (ACC) or Trimble Connect is almost unimaginable now. ACC even has an Australian data centre for lower latency. The dreadful ‘Revit Server’ is just a distant memory (although I believe it still exists) & I can work on cloud models at home in a sleepy river town
So, I’ll give myself a pass on this one
Databases & analytics 
In both my predictions and analytics posts from 2016, I foresaw the rise of analytics using model data. I saw a large part of this was not just retrospective analytics, but predictive. I didn’t see the potential of Machine Learning or Artificial Intelligence at that stage, but it was clear that the real value was predicting the future based on past data.
Just about everyone is producing PowerBI dashboards from their models now- some are claiming it as new & exciting.
So, I’ll give myself a big distinction on this one
Predictive modelling
I think I was on the right track here- I talked about Monte Carlo (probability analysis) without realising that a machine learning model, trained on historical data would be far superior. However, I was on the right track, although it hasn’t become commonplace just yet.
So, I’ll give myself a pass on this one
Machine control systems and robotics 
I correctly predicted that 3D printing techniques in construction were way off in the distance, but thought that a more direct link between model data and machine control or robotics systems would be in place by now. In other words, a machine would make something directly using the model.
Apart from a few token robot dogs for a photo-op & CNC machines, there has not been any meaningful robotics development in construction.
So, I’ll give myself a fail on this one.
Prefabrication
Prefab is a long-vaunted promise for construction- it’s simple in theory but very difficult in practice. In my blog, I recognised that it wasn’t as easy as making a whole building unit in a factory, chucking it on a truck and crane-ing it into place. But I did expect that there might be more in the way of prefab components.
So, I’ll give myself a fail on this one.
Internet of things (IOT)
I correctly predicted the rise of IOT sensors and devices, although still fairly limited within construction- but fairly commonplace in building management systems. I can’t recall if ‘Digital Twin’ was a thing at that stage, but I think I was on the right track.
So, I’ll give myself a pass on this one
Convergence
Although still a little separate, there has been a definite convergence of previously disparate things like ‘BIM’ and ‘GIS’. Some applications such as CityEngine are based on procedural modelling and could be considered as both BIM & GIS.
So, I’ll give myself a pass on this one
Competition
Since 2016, not much has changed in terms of competition- if anything, it’s probably got worse, with the incumbents consolidating market share. Genie BIM and the entire Google Construction Technology unit is long gone. I mentioned the existence of software resellers- I’m surprised they are still around and was wrong in thinking that they’d vanish.
So, I’ll give myself a fail on this one.
Software licencing
I correctly predicted the rise of usage-based cost models and that perpetual licences would be thing of the past. This now applies across just about everything, including Google Cloud, Amazon Web Services, Azure and databases such as Snowflake. Some BIM applications from Autodesk & others have gone part-way, being based on tokens or other usage-based models.
So, I’ll give myself a pass on this one
Note to McNeel: I still love Rhino and Grasshopper- apart from being fantastic applications, your licence/cost model is wonderful- please don't change a thing.
Data flow
I don’t think I was entirely correct, or entirely wrong on this. With the advent of ‘Digital Twins’, there is some data coming back into models from IOT devices.
So, I’ll give myself a sneaky pass on this one
Computational Design
Computational design applications such as Dynamo have gone from strength to strength, and now include applications other than just Revit. In addition, applications such as Forma are starting to include AI-based features (although I’d have to say it is very rudimentary and just scratching the surface.)
Machine Learning algorithms to generate 3D models are in their infancy, although you can see the potential. For example, point-e, stable-3d, GET3D can generate a vase of flowers, but not a structure or a building.
So, I’ll give myself a distinction on this one
Data Capture
I talked about LIDAR and photogrammetry, which although not brand-new were not commonplace.
Since then, there have been a lot of advances, particularly with cloud-based systems such as sensori or pointerra. Furthermore, desktop applications like Autodesk Recap and the use of pointclouds with Revit/Autodesk Construction Cloud are standard practice.
Although not the same, techniques such as Neural Radiance Fields (NeRF) offer exciting opportunities in data capture, visualisation & presentation.
However, I probably didn’t get it quite right with my dream of a more real-time process.
So, I’ll give myself a pass on this one
Summary
The natural next step is to put together a new list of technology predictions- I’ll get onto this….
In the meantime, the following post specifically covers AI opportunities in Construction.
Andrew Hannell: POV 
You must be logged in to post a comment.