In class last week we watched part of the youtube video, "The Future of BIM Will Not Be BIM—and It's Coming Faster than You Think”, in which Bill Allen discusses the current state of BIM, and where he believes the industry is heading in terms of automation and accessibility. One of the main questions he presents to the audience is, “ Can algorithms design a building?” He believes more and more work will be performed by computers and machines, and instead of building information modeling, we are going to see Building Information Optimization.
Mr. Allen explains the use of Building Information Optimization to expedite the design process by “feeding” the computer rules or constraints in design that have to be met, and the computer will construct the building’s optimal footprint, structural load capacity, and thermal performance.
I find the concept of algorithms that determine the optimal footprint of buildings extremely interesting, practical, and applicable to the entire industry, but it seems to me that it could remove an element of the human creativity process that is present in building design. While obviously these features will speed up the iterative design process, will the algorithms have too much control over the design that buildings begin to be repetitive to achieve streamlined efficient buildings? On the flipside of this argument, the algorithms could potentially create designs that an individual would have never had thought of. I believe the use of algorithms will be best used as a starting point where numerous potential designs are output by the algorithm, and then the design can be manually edited.
Allen believes that for the less niche designs, algorithmic building can be used effectively since the algorithms can mimic human behavior enough that the design options created will seemingly have no difference to designs manually created by humans. Additionally, with algorithmic design the number of design options output will be much greater than the number of outputs from humans while taking considerably less time.
In a more intermediate and current example, Allen brings up an example of how he has used graphical algorithmic modeling with Grasshopper to develop a family with complicated planar panels, in a very intuitive and graphically accessible way that does not require complicated coding or equation inputs. He refers to this creation as the “transition period”, before the computer algorithms will develop all of the design iterations on their own. I think graphical algorithmic modeling has more of a place in the coming years, than complete algorithmic modeling, where the computer creates the options for you, especially in the 3 to 10 year time frame which Allen proposes (this video is already more than 2 years old).
Source:
Allen, Bill . YouTube, The Future of BIM Will Not Be BIM—and It's Coming Faster than You Think, 23 Nov. 2016, www.youtube.com/watch?v=xq6yKyauu-o.
Comments:
1) Nick,
Mr. Allen explains the use of Building Information Optimization to expedite the design process by “feeding” the computer rules or constraints in design that have to be met, and the computer will construct the building’s optimal footprint, structural load capacity, and thermal performance.
I find the concept of algorithms that determine the optimal footprint of buildings extremely interesting, practical, and applicable to the entire industry, but it seems to me that it could remove an element of the human creativity process that is present in building design. While obviously these features will speed up the iterative design process, will the algorithms have too much control over the design that buildings begin to be repetitive to achieve streamlined efficient buildings? On the flipside of this argument, the algorithms could potentially create designs that an individual would have never had thought of. I believe the use of algorithms will be best used as a starting point where numerous potential designs are output by the algorithm, and then the design can be manually edited.
Allen believes that for the less niche designs, algorithmic building can be used effectively since the algorithms can mimic human behavior enough that the design options created will seemingly have no difference to designs manually created by humans. Additionally, with algorithmic design the number of design options output will be much greater than the number of outputs from humans while taking considerably less time.
In a more intermediate and current example, Allen brings up an example of how he has used graphical algorithmic modeling with Grasshopper to develop a family with complicated planar panels, in a very intuitive and graphically accessible way that does not require complicated coding or equation inputs. He refers to this creation as the “transition period”, before the computer algorithms will develop all of the design iterations on their own. I think graphical algorithmic modeling has more of a place in the coming years, than complete algorithmic modeling, where the computer creates the options for you, especially in the 3 to 10 year time frame which Allen proposes (this video is already more than 2 years old).
Source:
Allen, Bill . YouTube, The Future of BIM Will Not Be BIM—and It's Coming Faster than You Think, 23 Nov. 2016, www.youtube.com/watch?v=xq6yKyauu-o.
Comments:
1) Nick,
I agree that the algorithmic tools presented in the lecture have great potential in the speed and quantity of iterative design. It also seems to be applicable to any discipline so it will have a universal use in the AEC industry. To me, AI BIM seems to most functional and appropriate when the designs work cooperatively between the AI and human input. I think the biggest hurdle in the adoption of these programs is their cost, learning curve and accessibility for a slow adopting industry like the AEC.
2) Albert,
I agree Allen’s prediction for the state of AEC tech after BIM is unrealistic. In the video he says 3 to 10 years, and it is already 2 years since he gave that presentation. Interoperability and accessibility are huge hurdles for the industry, and it seems that unless both of these are very refined for these algorithm modeling programs, they will not be implemented in a practical way anytime any time soon. I appreciate your personal insight into the speed at which firms adopt new technology. The firms I worked at all primarily focused on Revit, but the idea that some companies are dragging their feet on Revit makes Allen’s prediction seem almost impossible for the majority of the industry in 10 years.
3) Weiyi,
Good points on the speed of implementation for software and technology, as well as reasons why technology sometimes seems to fade into and out of relevance. I agree that AI assistance can be a great boon for the AEC industry in terms of design optimization, as long as humans are making the final choice in the design. I think it will be a little hard to market completely computer design buildings to the general public for a while. The potential of this technology is remarkable. AI BIM can lead to increased efficiency and design speeds, and with that more time will be available for individuals to learn new skills or do other things for projects. It could completely change the way and quality of life for individuals in the AEC industry.
Sean, It is good to see that you are taking an interest in the algorithms mentioned by Bill Allen because they will in no doubt become relevant at some point in your career. I find the drawbacks to the use of design in repetitive buildings being an interesting limiter to the capabilities of these database programs. I still personally believe that 10 years may be a bit soon for all of these programming changes to transpire.
ReplyDelete