“The built environment is an $8 trillion per year industry that is still basically artisanal.” So said Astro Teller, head of the Google X research lab, during a speech at South by Southwest last week. Reading that sentence in isolation, you might assume that Teller intended it as praise, that he was was applauding the field of architecture for maintaining its heritage of craftsmanship, skill, and artistry. But you would be wrong. Being “still basically artisanal” is, for Teller, a great flaw. It’s a symptom of both a debilitating lack of software-mediated routinization and a tragic superfluity of quirky human talent. Artisanality is a problem that Google is seeking to solve. One Google X project, Teller explained, is intended “to fix the way buildings are designed and built by building, basically, an expert system, a software Genie if you will, that could take your needs for the building and design the building for you.” By getting all those messy and outmoded artisans out of the picture, replacing them with tidy software algorithms, we’ll be able to avoid the inefficiency and waste that inevitably accompany human effort.
But, Teller went on to say, the Genie project has run into a problem: “We found out that the system we envisioned couldn’t connect into the infrastructure and ecosystems for building the built environment because that software infrastructure is piecemeal and often not software at all but just knowledge trapped in the heads of the experts in the field.” Let me repeat that last bit: “not software at all but just knowledge trapped in the heads of the experts in the field.” Quelle horreur! The goal now, he said, is to take “a huge step back” and lay “a software foundation and data layer” that will allow Google to liberate all that head-imprisoned knowledge and eradicate the pestilence of artistry once and for all.
Image: Mark Moz.
Quelle horreur indeed! As a translator, I hear versions of this story in the form of how machine translation (aka, Google Translate and similar engines) will “solve” the translation problem in many organizations. Translated texts are now “translated content” and everybody in the chain is a “knowledge worker,” a designation I find both puzzling and disgusting.
If we could only convey the message that no software can ever think because only humans think, we could be a step ahead.
How muddleheaded can a person be? I nominal Astro Teller (what a name …) as Chief Muddlehead. Or maybe it should be Chief Hammerhead, since he seems to believe that all the thorny problems formerly addressed by human know-how are now just nails to be pounded down by software.
Thank you Nick for sharpening your rhetoric about the Bizarro world these soulless ghouls want to foist on us. Every one of us in the IT industry needs to know where his automation frontier lies. Reading your books has helped me say thus far and no further.
You have laid an solid historical and intellectual groundwork and I encourage you to hoist the appropriate people on the appropriate petards.
There is a case to be made for lowering housing costs using mass production techniques like prefab housing made in factories. They have little to nothing to do with software. For an illustration, see how China Vanke builds entire apartment buildings in days using prefabricated elements. Fancy and inspiring it may not be, but it provides affordable housing to China’s rapidly urbanizing masses, a worthy endeavor.
An older example is Eichler and his affordable but high-quality midcentury modern houses for the working class (which included African Americans, he was one of the few builders not to discriminate against them). They have a cult following nowadays.
Google’s central goal is AI.
The article is rather nostalgic: We humans are the best and will be the best forever in all tasks … This is (unfortunately?) not true.
There is a sequence of human tasks of increasing complexity cp. to increased market value; for a computer company, it is completely legal to investigate the technical and business potential of assistance or replacement of human activities by computers. Where would we be without this attitude? No weaving looms, no text software. Isn’t it just fair and great that one of the hardest activities for computerization is the blue collar work? At least in such a flexible and fuzzy environment as the built environment.
But to be clear: This is just an example that at the end only arts remain unique for humans, and even there it is not hard-wired what art really is.
A funny thought: in a few years, Google will be sitting on the ashes of so many of its crazy AI projects, having burned billions of dollars in advancing the unattainable.
Can software mimic some human activities or behaviors, or even animal ones? Sure, but instead of building an AI entity worth millions of dollars, why not hire a helper, a nurse or just buy a cat?
Same goes for IBM: millions poured into Watson. Do you know what Watson is doing these days? Its helping cooks.
I applaud the world’s artisans, whether they’re potters, architects, writers or watchmakers. You will never be truly replaced.
I don’t understand the snark. “Knowledge trapped in the heads of experts” has always been a problem. This is why we have books, scientific articles, and manuals. To make it available to anyone interested.
Wow. Yes! Extract the knowledge and promptly relay it here! I didnt know it was possible to make ourselves more obsolete. I mean, thats what were working toward, right?
Wish I’d seen this talk at SXSW this time around. There’s such a long history of AI and expert systems in architecture that goes back to really the beginnings of AI, with all of its DARPA and Office of Naval Research funding. Architecture just doesn’t solve all that easily, it turns out…
Could it be that Google wants to be the only artist? The one? The big guy?
I am reminded of the introduction to “Algorithms for Visual Design Using the Processing Language” by Kostas Terzidis.
“It is possible to claim that creativity is limited by the very programs that are supposed to free their imagination. … …Whenever they use a new tool provided for them by programmers, they think that they are now able to do something new and “cool.” But are they really doing anything new? Or are they simply replicating a process already conceived by the programmer who provided the tool?”
That’s a thought-provoking quote, Timothy. Perhaps every course on software development should have an intro to Humanities (not an elective, but a mandatory course) so that students can learn and understand the limits of technology. A chronology of promising but failed technologies would be a useful reminder as well.
About buildings/architecture/urbanism, and maths/science, it is quite refreshing to read the works of Lionel Marsh and Leslie Martin (in Cambridge in the sixties), see for instance the pdfs linked below (blog post in French but pdfs in English) :
https://iiscn.wordpress.com/2011/05/15/densite-etages-lumiere/