Log in

goodpods headphones icon

To access all our features

Open the Goodpods app
Close icon
Moore's Lobby: Where engineers talk all about circuits - Getting Out of the Weeds with John Deere’s Revolutionary AI Technology

Getting Out of the Weeds with John Deere’s Revolutionary AI Technology

06/13/23 • 55 min

Moore's Lobby: Where engineers talk all about circuits

Imagine 1500 laptops strapped to a 120 foot (36 m) wide sprayer being pulled behind a tractor. That gives you a rough understanding of the compute power inside the John Deere See & SprayTM that targets herbicide to the weeds, not the crops. These might be “the smartest agricultural machines in the world,” says Chris Padwick, Director of Computer Vision and Machine Learning at Blue River Technology, a wholly owned subsidiary of John Deere.

Padwick was not looking to return to his agricultural roots, but in a “shields down” moment” he was approached by a startup, Blue River Technology. He recalls being “completely blown away” by what they were doing in computer vision.

After initially developing a product for lettuce thinning, Blue River pivoted to the problem of row crop weeding. In the US alone, the agricultural industry uses over 20 million gallons of herbicide annually, almost all of it through broadcast spreading across the entire field. Blue River set to work on changing the paradigm to only spray the weeds. According to Padwick, this “dot matrix printer for weeds” has the potential to “save 15 million gallons of herbicide” and “about a billion dollars.”

In this episode of the Moore’s Lobby podcast, Padwick takes us behind the scenes to discuss the challenges of data collection in the fields, image classification, the compute technology that drives these real-time systems, and much more. On this business side, he reflects on the 2017 acquisition by John Deere and how important that was in their development of products at scale that can operate in rugged environments for two decades.

“I feel like we're going to look back on farming in five or ten years and we're just going to be amazed...Precision spot spraying is going to be the way that people treat their crops going forward because it's going to allow them to save so much more chemical. And the sustainability in the environment is kind of unparalleled. So I really feel we're kind of at that inflection point for agriculture.”

So, listen in on this fun interview as Padwick describes:

His graduate research into Big Bang cosmology and background radiation

Why he left engineering as an undergrad—seriously, universities we have to fix this outdated solution to engineering education!

Blue River’s global operations to collect crop and weed images to drive continuous improvement

plus icon
bookmark

Imagine 1500 laptops strapped to a 120 foot (36 m) wide sprayer being pulled behind a tractor. That gives you a rough understanding of the compute power inside the John Deere See & SprayTM that targets herbicide to the weeds, not the crops. These might be “the smartest agricultural machines in the world,” says Chris Padwick, Director of Computer Vision and Machine Learning at Blue River Technology, a wholly owned subsidiary of John Deere.

Padwick was not looking to return to his agricultural roots, but in a “shields down” moment” he was approached by a startup, Blue River Technology. He recalls being “completely blown away” by what they were doing in computer vision.

After initially developing a product for lettuce thinning, Blue River pivoted to the problem of row crop weeding. In the US alone, the agricultural industry uses over 20 million gallons of herbicide annually, almost all of it through broadcast spreading across the entire field. Blue River set to work on changing the paradigm to only spray the weeds. According to Padwick, this “dot matrix printer for weeds” has the potential to “save 15 million gallons of herbicide” and “about a billion dollars.”

In this episode of the Moore’s Lobby podcast, Padwick takes us behind the scenes to discuss the challenges of data collection in the fields, image classification, the compute technology that drives these real-time systems, and much more. On this business side, he reflects on the 2017 acquisition by John Deere and how important that was in their development of products at scale that can operate in rugged environments for two decades.

“I feel like we're going to look back on farming in five or ten years and we're just going to be amazed...Precision spot spraying is going to be the way that people treat their crops going forward because it's going to allow them to save so much more chemical. And the sustainability in the environment is kind of unparalleled. So I really feel we're kind of at that inflection point for agriculture.”

So, listen in on this fun interview as Padwick describes:

His graduate research into Big Bang cosmology and background radiation

Why he left engineering as an undergrad—seriously, universities we have to fix this outdated solution to engineering education!

Blue River’s global operations to collect crop and weed images to drive continuous improvement

Previous Episode

undefined - VC-Backed Hardware Design Collaboration is the (All) Spice of Life

VC-Backed Hardware Design Collaboration is the (All) Spice of Life

In this unique episode of the Moore’s Lobby podcast, we get to hear from not only the technologists and executives behind the company but the investor who is helping them make their dreams a reality.

As an electrical engineer, Chrissy Meyer had spent years working in product design on large projects at Apple, including the Apple iPod Nano and Touch. She is all too familiar with the outdated methods of hardware collaboration and design reviews—printed schematics, highlighters, and engineers huddled around a conference room table.

According to Kyle Dumont, a first-time entrepreneur as Allspice CTO, their goal is centered “around making it easy to build a workflow and collaboration process for electronics designs.” In his earlier career working in hardware at both large and small companies, he also “became pretty frustrated with how difficult it was to collaborate on our electronics designs.”

Valentina Ratner, CEO of Allspice, goes on to explain that the industry is “trying to build colonies in Mars, and I don't think we're going to get there with Google Sheets and PDFs and screenshots.”

In her current role as a partner at Root Ventures investing in early-seed tech startups, the engineer inside Meyer had a “visceral” reaction when Ratner and Dumont explained their vision to bring Git-style collaboration software to the hardware world. Meyer explains that “there was never a doubt in my mind that this problem that they were describing was very real. Because the truth is I had lived it for 15 years.”

Meyer said, “I had to take off my engineering hat for a quick second and say, okay, reign in the excitement. Yes, this is a fantastic product for an engineer. Is this a great investment?...the truth is, there are far fewer electrical engineers than there are software developers.”

With the backing of Root Ventures, Ratner and Dumont founded Allspice and quickly got to work. As a first-time entrepreneur and CEO, Ratner admits that she had to learn a lot quickly because “there are just so many unknowns.”

Reflecting on some early lessons, Ratner shares some interesting advice for fellow tech entrepreneurs:

“It's less about convincing anyone and more about finding the people that already believe in the version of the world that you're trying to build.”

Other highlights of this round-table discussion include:

-Meyers’ belief that “'there's this common misconception that in order to approach VCs, you need to have a fully polished pitch deck, and you need a financial model and forecasts and projections.”

-Insights on one of the favorite features of Allspice: the ability to run digital, asynchronous hardware design reviews

-Where Meyer thinks hardware and software companies are going to win in the future

Next Episode

undefined - AI or 5G? EdgeQ Asks, “Why Not Both?” While Building a Wireless Base Station-on-a-Chip

AI or 5G? EdgeQ Asks, “Why Not Both?” While Building a Wireless Base Station-on-a-Chip

It all began with a love of mathematics. Adil Kidwai loved mathematics. That initially led him to a career focused on analog and RF design. After many years working on RF technology leading efforts in cellular communication, Wi-Fi, and Bluetooth, Adil had a chance to move into the AI field.

When he started digging into AI, Kidwai had an epiphany. “I realized that AI borrowed a lot of concepts from information theory which were developed 50 years ago.” The mathematics of RF communication and artificial intelligence were basically identical: convolutions, matrix multiplication, nonlinear operations, and more.

It was during this time that Kidwai met EdgeQ CEO and founder Vinay Ravuri for lunch, and it “just clicked”. Kidwai recalls that they were thinking exactly the same things about the relationships between AI and 5G. So, Kidwai joined EdgeQ, where they have developed an integrated circuit that leverages the customization capabilities of RISC-V to enable and optimize both AI and 5G.

They believe that their “software-defined base station-on-a-chip” allows AI to complement 5G in some applications and 5G to complement AI in others. As Kidwai notes, “these two technologies support each other in the world that we are living in.” And, they already have their eye on 6G.

In this Moore’s Lobby interview with our host Daniel Bogdanoff, Kidwai shares how different it was when moving from a behemoth like Intel to a startup. “You have to move like 100 mph on day one!...because it’s a question of existence all of a sudden.”

Here are a few other highlights from this interview with Kidwai:

-Think your master’s thesis was difficult? Wait until you hear what Kidwai had to accomplish at UCLA.

-His reflections on spending several years in Munich working on Intel’s acquisition of Infineon’s wireless group. He even has some recommendations for people considering moving abroad.

-The surprising changes in the wireless world Kidwai has seen in his decades working in the industry.

There is much more, so go listen for yourself! Please tell us what you think in the comments. We would also love to hear your recommendations for the companies and people you would like to hear from in future episodes of Moore’s Lobby.

Episode Comments

Generate a badge

Get a badge for your website that links back to this episode

Select type & size
Open dropdown icon
share badge image

<a href="https://goodpods.com/podcasts/moores-lobby-where-engineers-talk-all-about-circuits-114429/getting-out-of-the-weeds-with-john-deeres-revolutionary-ai-technology-30750667"> <img src="https://storage.googleapis.com/goodpods-images-bucket/badges/generic-badge-1.svg" alt="listen to getting out of the weeds with john deere’s revolutionary ai technology on goodpods" style="width: 225px" /> </a>

Copy