INSIDER: Hypar Looks to Unlock AEC Industry Expertise via Computational Design Ecosystem

DO YOU HAVE INDUSTRY EXPERTISE? If you have been in the AECO (architecture, engineering, construction, operations) industry long enough, the answer is “yes.” Yet, you may not realize you have more expertise that is convertible to computational design methods than you understand.

Naturally, you may wonder how? How does this expertise get formalized into a digital tool that you can use, that others can use, particularly one that is optimized at multiple levels of use, deployment, speed, and power?

This is where Hypar comes in. The company is setting up an AAD (algorithms-aided design) platform and ecosystem that aims to help AECO companies unlock, democratize, and deploy their expertise—at scale! These companies range from architecture and engineering firms, real estate and development professionals, to folks on the manufacturing and design-to-fabrication side of the industry. No matter where you sit in the industry, you have “pain points” that your side of the industry has been living with since you entered it. Who in the industry doesn’t wish to change that?

 

 

 

 

Well, it turns out many do. And folks are tapping the power of computational design tools, or AAD tools, to transform their workflows and solve age-old problems at multiple factors faster than previous methods or embed their hard-won specialized expertise, often enshrined in Excel spreadsheets, into 3D visuals and KPI’s they often hold close to their chest.

Hypar’s Mission

Hypar’s founders believe that after 5000 years of building the world, no AECO project should start from a blank page. With that in mind, they have created a web-based platform for generating buildings using encoded expertise in the form of algorithms-aided design (AAD) functions. You may ask: who is providing the basic functions?

The Hypar development team wrote a core group of functions called “common” functions. All other functions are written by developers who are “users” of the Hypar platform. Basic functions in the common set include things like fitting a building to a given site context, determining that building’s shell and core design, and essential facade generator tools and more.

So far, industry reaction to Hypar has been measured yet excited. This makes sense; AAD tools in the AEC industry are still adopted by architecture firms in the single-digit percentages of all practices. Still, reaction at Autodesk University, where Architosh last saw Hypar, was very solid with many curious to understand what this cool-sounding company was all about.

How Hypar Works

So how exactly does Hypar work?

First and foremost, this is a system for distributed (SaaS-based) computational design. It means through the web browser folks can solve problems—a whole array of them—through the Hypar platform and ecosystem of solutions.

Hypar’s AAD-based functions come in three classes. The “common functions” are available to all monthly subscribers. Think about the new Disney+ video service where you get all the Disney classics as part of your set monthly subscription. Same idea. These common functions are supplied by Hypar and by any other “contributors” who choose to make their functions available to all Hypar subscribers.

Another class of functions is “your functions,” the ones your company has crafted exclusively for their use but distributed through the Hypar platform. These functions could be developed by in-house developers, external developers, or the Hypar development team, but they encapsulate your expertise reignited through new computational design tools.

Finally, the third class of functions is “those written by software firms” who have chosen Hypar as a pathway to market. “We have had several software development companies talking to us,” says Anthony Hauck, “for a while folks like TestFit have experimented with the platform, just to find out what it would be like to put their building shell generator on Hypar.”

Hauck has said, Hypar is one approach to the market. “Hypar isn’t just for people who want to put together lists of functions and make something from them, or even just for people who want to write functions from scratch, but also for software companies interested in going to the Web on a very easy path.” For that, Hypar has its own set of free development tools, so developers and scripters of AAD solutions can get started quite easily.

But Why Hypar?

There are many arguments for the Hypar proposition, but here is one that may make the most sense to the largest group of AECO users if they could stop being so overcautious about it: deploy one’s expertise by encoding it into “computational design” and decision support tools.

“We see a lot of organizations trying to normalize and distribute their expertise internally,” says Hauck, “a lot of times the vehicle is simply Excel. One or two experts will record that expertise in a spreadsheet, and they will distribute it by either sending a link to it on the network or by email—and you can imagine the update nightmare that attends that procedure—but that’s what they do.”

Real estate development professionals, in particular, have decades of proprietary expertise, often localized to various markets or client group types. While the numbers help them with key decision support, there is often zero graphical representation 2D or 3D that attends this data. This is a natural shortcoming of that is easily fixable, says Hauck. There are already specific AAD tools sprouting up that address the pre-design stages of the building industry. Still, those tools don’t necessarily allow professionals and companies to encode their specific expertise in a way that keeps that expertise private. Hauck says that Hypar never gets any source code at all. “If someone shares a function to the Hypar platform, what we get is a compiled DLL.”

As readers know, some AECO companies already have their computational design professionals creating bespoke tools. “I’m talking to an engineer about the use of our platform, and he says they are very good at engineering big box stores,” says Hauck. “And what that means is they tie up a PE on a big box store that is nearly exactly the same as all the other big box stores they do, even if for multiple clients. So they are looking at how do they automate MEP design tasks so their PEs can free up time for more interesting and often more lucrative work?”

At Autodesk University 2019, there was a session on Hypar by an MEP firm that was using it for its engineering workflows. This is the type of firm that is tired of not doing anything about age-old pain points in existent AEC workflows and has finally decided to invest the time to do something about it.

Hypar’s Ecosystem Argument

Hypar is both a platform and an ecosystem. One side of the Hypar equation is about using AAD-based functions as an end-user, solving your building industry problems. The other side is about uploading your AAD-based functions for distribution to others. “Others” can be your company colleagues, business partners, clients, or anybody else on the Hypar platform. The system is designed to encourage sharing.

“We have a structure that encourages openness among contributors,” says Hauck. “If you want to expose your content to everyone who is a subscriber on Hypar, and you expose your code on Github, everything is free to you. Anywhere in between, you are paying for various degrees for privacy.” Hauck and his co-founder Ian Keough haven’t yet finalized their pricing model, but Hauck describes it has a mashup of Amazon Prime, plus Github, plus the Apple App Store.

If you are a big company that wants to use Hypar to distribute your encoded expertise in computational design—say to your global offices—but not expose your code on Github, the fee may be USD 10 per month to host five functions. And Hauck says it scales from there, but “things get cheaper if you share more.” The goal is to get companies and individuals to share their functions, creating a vast ecosystem of targeted computational design tools for the benefit of the entire AECO community.

For software firms and others—like industry-leading three-letter architecture firms—who want to share their computational design expertise for a fee, Hypar is developing a financial model as well. “Suppose somebody comes up with a much better facade creator than the “common” function for that,” adds Hauck, “if they want to charge a per-use or subscription fee for that, we will be supportive of that model also.”

Hypar—Looking Forward

The Hypar vision is compelling as a democratized vision, but industry incumbents may be wary. Today computational design use currently resides in multiple leading AAD toolchains such as Rhino + Grasshopper, Bentley’s Generative Components tied into OpenBuildings Designer, or Marionette tied in Vectorworks Architect, or Dynamo and its strong connection to Revit. Part of Hypar’s challenge is not just competing with those tools but convincing those who already hold encoded expertise across those platforms to place their algorithmic-aided design (AAD) functions on Hypar, to bring them into the world as web services.

 

 

 

 

To get there, Hypar needs to support more forms of encoded expertise. “We see in our industry, five or maybe six areas where repositories of building expertise reside,” says Hauck. He is specifically talking about this expertise residing inside codes or scripts developed in different languages. They include in alphabetical order: C#, Dynamo, Excel, Grasshopper, Python, and Revit families. Hauck says, “We intend to make all of those a part of the Hypar ecosystem in the future.”

For now, in the immediate roadmap, Excel contains a massive amount of encoded industry expertise, and Hauck says they intend to link Hypar to Excel in the very near future. By linking in Excel, Hypar can target a lot of expertise in the pre-design areas of the building industry, as well as a vast array of industry knowledge in building typology planning expertise.

As for the grand scale potential of industry expertise? Hauck says he is talking to a lot of CEOs and telling them that the market for that expertise is tremendous and that few folks can tap that expertise. The vast majority of architects do not have computational design capacities of any kind, nor do they realize how to use such tools for their problems. Yet, in most cases, they are sitting on expertise nonetheless.