There is a certain irony in an industry capable of erecting multibillion-dollar semiconductor plants, nuclear reactors and sprawling data centers, but still struggling to harness the efficiency of waste contained in 40-yard dumpsters.
This contradiction is at the heart Hungry for innovationposted on October 1 by Todd Thomas, co-founder and CEO of construction waste technology company Woodchuck.
The book argues that some of the most important technological advances can come not from inventing something new, but from applying existing tools to problems that companies have learned to tolerate.
Thomas brings an unusual perspective. While leading an in-house technology company within toolmaker Stanley Black & Decker’s Radical Innovation Accelerator, he helped develop a connected drill designed to capture installation and torque data, potentially streamlining inspections. The prototype never made it to market, which Thomas attributes in part to corporate resistance, though the book also acknowledges the disruptions of the pandemic era.
This experience helps explain the more pragmatic approach the author later took at Woodchuck, based in Grand Rapids, Michigan. Instead of inventing another technology, the company applies existing AI image recognition tools to identify discarded building materials, reclaiming wood to transform into biomass fuel for electricity generation.
The connection between construction waste and energy production also informs the book’s wider ambitions.
Thomas examines nuclear power, renewables and AI’s growing demands for electricity, situating Woodchuck’s business model within a wider discussion of energy supply. His treatment of these technologies is more introductory than exhaustive, but the book is most persuasive when it draws on his first-hand experience in overcoming the practical obstacles to bringing an innovation to market.
The strongest chapters follow Woodchuck’s development from an idea to reclaim discarded construction wood in an AI-assisted waste management operation. Thomas describes the company’s early work on a Ryan Cos. project. for Amazon, where its image recognition technology failed to identify materials covered by an overnight snowfall.
Looking for quick answers on construction and engineering topics?
Try Ask ENR, our new intelligent AI search tool.
Ask ENR →
Thomas writes conversationally, with a storyteller’s eye for the setbacks and personalities that make innovation unpredictable. His enthusiasm sometimes morphs into motivational business prose, but the book shines when it eschews sweeping predictions and instead describes how his ideas fit the realities of construction.
For contractors, project managers, and construction technology developers, these accounts offer the clearest value in the book. Thomas’ emphasis on financial returns, operational simplicity and customer adoption is familiar territory, but first-hand examples give substance to these principles.
His most useful insight is also his least grandiose: innovation succeeds when it solves a problem that people really need to solve.
In an Oct. 9 interview with ENR, Thomas discussed the industry’s skepticism toward new technologies, the economics of waste recovery and the lessons learned from Woodchuck’s early mishaps.
This interview has been edited for length and clarity.
***
ENR: Is construction really hungry for innovation or is it appropriately skeptical of unproven technology?
Thomas: Honestly, it’s probably a combination of both. … People generally resist change. In the construction industry in particular, I mean, there are a lot of old pros who have been doing the exact same job with the same subs for 30 years, so they do it the way they do it, and that’s what they know, and that’s what they like.
Asking them to change is hard, and to your point of healthy skepticism, I mean, a lot of people have introduced different technologies over the years and tried to make improvements, and a lot of them didn’t work, so I think you’re right.
His skepticism is appropriate.
ENR | What separates useful build AI from AI hype?
Thomas: Lots and lots of people are using AI just for the sake of using AI. … I think the approach should be the opposite. Start with the problem you’re trying to solve. Let the problem define the solution. In our circumstances, AI image recognition is a perfect technology to help us identify and classify different materials so that we can divert them and keep valuable materials like wood and cardboard and plastic and metal out of the waste stream.
But many problems don’t need AI. So don’t use AI just for the sake of using AI. Use the right tool for your problem.
ENR | How does Woodchuck’s approach reduce construction waste costs?
Thomas: What Woodchuck does is separate these materials on site and consolidate them. … We take all those pallets and send them to a grinder and send them [the wasre] over a magnet to remove all the metal and thus collect all the nails, staples and screws, and recycle that too.
But then we have this huge pile of wood chips and we load it into a semi-truck. We can reach 25 to 30 tons in the semi-truck. So you can imagine the efficiencies there. Instead of hauling two tons at a time off site, we are now hauling 25 tons at a time off site. The other part is when we take these wood chips [and] those pallets… and turn them into wood chips, we call it biomass. At this point, we sell it to the local energy company that uses it to produce electricity.
ENR | What has Woodchuck learned by failing?
Thomas: Every start fails. We certainly have our list of failures. … It was January when we launched this first project. We built and deployed our image recognition system, and we were very excited on the first day to look at… the data we collected through our cameras, and it was all crap. It didn’t make any sense. We’re like, what the hell?
It had snowed that night. All the wood was covered with snow, and our system didn’t know what snow was. We couldn’t see anything, so the first day was a complete failure. But we learned a very good lesson and our response was that we built a full-size permanent facility in Grand Rapids.
Each container 1791831302 it has two cameras, so we’re capturing images all day, every day, in 10 bins, and we rotate the different materials and different contaminants.
So we’re always teaching our AI, and we’re teaching it 24/7, regardless of the weather.
