Episode 172: Systems, Software, and Stewardship with Matthew Fitzgerald

Matthew Fitzgerald is a “1.5 generation” organic grain farmer operating 2,500 acres in Minnesota, situated where the big woods meet the prairie. Alongside his father, Joe, he manages a diverse rotation of crops on land originally preserved by organic pioneer Mabel Brelia. Matthew focuses on navigating modern economic challenges while actively fostering opportunities for the next generation of land stewards.

To manage scale and improve yields, Matthew founded FarmFlow, a system that began as a visual whiteboard to track field passes and has evolved into software for optimizing execution. By integrating these standard operating procedures with emerging AI tools, he aims to professionalize organic operations and prove the financial competitiveness of sustainable agriculture.

In this episode, John and Matthew discuss:

Overcoming high barriers to entry for beginning farmers through creative partnerships and policy.
The concept of FarmFlow and using visual tools to manage complex organic systems.
Data analysis revealing the strong correlation between field pass frequency and increased crop yields.
Decommoditizing organic grain to ensure financial viability and economic independence.
Leveraging AI tools like FieldLark to enhance agronomic decision-making and operational efficiency.
The importance of team diversity and “quorum sensing” in building resilient farm operations.

Additional Resources
To learn more about FarmFlow, please visit: https://www.tryfarmflow.com/

To learn more about Matthew, please visit: https://www.fitzgeraldorganics.net/

About John Kempf
John Kempf is the founder of Advancing Eco Agriculture (AEA). A top expert in biological and regenerative farming, John founded AEA in 2006 to help fellow farmers by providing the education, tools, and strategies that will have a global effect on the food supply and those who grow it.

Through intense study and the knowledge gleaned from many industry leaders, John is building a comprehensive systems-based approach to plant nutrition – a system solidly based on the sciences of plant physiology, mineral nutrition, and soil microbiology.

Support For This Show & Helping You Grow
Since 2006, AEA has been on a mission to help growers become more resilient, efficient, and profitable with regenerative agriculture.

AEA works directly with growers to apply its unique line of liquid mineral crop nutrition products and biological inoculants. Informed by cutting-edge plant and soil data-gathering techniques, AEA’s science-based programs empower farm operations to meet the crop quality markers that matter the most.

AEA has created real and lasting change on millions of acres with its products and data-driven services by working hand-in-hand with growers to produce healthier soil, stronger crops, and higher profits.

Beyond working on the ground with growers, AEA leads in regenerative agriculture media and education, producing and distributing the popular and highly-regarded Regenerative Agriculture Podcast, inspiring webinars, and other educational content that serve as go-to resources for growers worldwide.

Learn more about AEA’s regenerative programs and products: https://www.advancingecoag.com

Podcast Transcript

0:03 – 0:04

Hi friends, this is John.

0:04 – 0:05

Welcome back to the Regenerative

0:05 – 0:06

Agriculture podcast, where we

0:06 – 0:09

have all kinds of fun

0:09 – 0:10

conversations related to

0:10 – 0:11

improving our

0:12 – 0:14

soil health and improving the

0:14 – 0:15

health of the things that we're

0:15 – 0:16

growing.

0:16 – 0:18

And in today's world, today's

0:19 – 0:20

macroeconomic climate,

0:21 – 0:22

perhaps one of the most

0:22 – 0:24

important conversations is how

0:24 – 0:26

do we maintain and improve the

0:26 – 0:27

financial health of our

0:27 – 0:28

operations? Because obviously

0:28 – 0:29

that is a foundational

0:29 – 0:30

prerequisite upon which

0:30 – 0:31

everything else rests.

0:31 – 0:32

Because if we don't have the

0:32 – 0:33

resources,

0:34 – 0:36

To be good stewards,

0:36 – 0:37

then we

0:39 – 0:39

all know what happens.

0:40 – 0:41

Stewardship suffers and declines

0:41 – 0:42

and the

0:43 – 0:44

things that we're responsible

0:44 – 0:47

for managing necessarily go into

0:47 – 0:48

decline.

0:49 – 0:51

And this has been a macro trend

0:51 – 0:53

for the last

0:54 – 0:56

century or more.

0:56 – 0:58

You could argue even longer that

0:58 – 0:59

there have been all kinds of

0:59 – 1:01

externalities in agriculture and

1:01 – 1:02

that

1:02 – 1:04

there has been this extractive,

1:04 – 1:06

this financial extraction that's

1:06 – 1:07

occurred from agriculture.

1:07 – 1:07

But this has really been

1:07 – 1:09

accelerated in the last 50 years

1:09 – 1:10

or so

1:10 – 1:11

and is continuing.

1:11 – 1:13

And so it's important for us as

1:13 – 1:15

stewards and as farmers to be

1:15 – 1:17

thinking about how do we put

1:17 – 1:18

ourself into a different

1:18 – 1:18

position.

1:19 – 1:20

How do we man and this is one of

1:20 – 1:21

the topics I'm going to be

1:21 – 1:22

speaking about next week here at

1:22 – 1:22

the acres conference.

1:23 – 1:23

But anyway,

1:24 – 1:25

too much of a monologue.

1:25 – 1:26

I'm delighted.

1:26 – 1:27

I'm joined here today by Matthew

1:27 – 1:28

Fitzgerald. Matthew, thank you

1:28 – 1:29

for joining me.

1:29 – 1:31

Tell us a little bit about your

1:31 – 1:32

story, your background, the

1:32 – 1:33

scope of the things that you're

1:33 – 1:34

working on on your farm in

1:34 – 1:35

Minnesota.

1:35 – 1:36

Yeah. Hey, John, thanks for

1:36 – 1:37

having me.

1:38 – 1:39

1:39 – 1:41

appreciate that opening comment

1:41 – 1:41

because

1:44 – 1:46

I'll tell our story and how it

1:46 – 1:47

connects to financial

1:47 – 1:48

stewardship and we'll see where

1:48 – 1:49

it goes from there.

1:51 – 1:52

I'm a second generation farmer.

1:53 – 1:54

My parents, Anna Jo,

1:55 – 1:57

moved to Minnesota from Iowa in

1:57 – 1:59

the year 2000, so 25 years ago.

2:01 – 2:03

We are a 100 % organic grain

2:03 – 2:03

farm.

2:04 – 2:06

I'll share a little bit about

2:06 – 2:08

the origin story of our farm and

2:08 – 2:10

how we as beginning farmers, I

2:10 – 2:12

would consider myself a 1 .5

2:12 – 2:13

generation because my folks

2:13 – 2:14

started in their 40s.

2:14 – 2:15

They were kind of a

2:16 – 2:18

midlife crisis, wanted to become

2:18 – 2:20

farmers, raise kids on a farm

2:20 – 2:22

context, but weren't

2:23 – 2:23

multi -generation.

2:24 – 2:25

So how did they get into it?

2:26 – 2:27

And really that was the organics

2:27 – 2:29

niche. And our farm has really

2:29 – 2:30

paralleled the growth of the

2:30 – 2:31

organic industry.

2:32 – 2:33

And now as a return to the farm,

2:34 – 2:35

thinking about this next

2:35 – 2:36

generation,

2:36 – 2:38

regenerative is kind of the next

2:38 – 2:39

thing that comes along, but

2:39 – 2:40

I'll, I'll give some more

2:40 – 2:42

background about our, our farm

2:42 – 2:43

and how it got started and what

2:43 – 2:44

we're up to today.

2:44 – 2:45

2:45 – 2:48

the farm was originally owned by

2:48 – 2:49

a woman named Mabel Brelia and

2:49 – 2:51

Mabel was an OBGYN doctor.

2:51 – 2:53

And she was really interested in

2:53 – 2:54

women and children's health.

2:54 – 2:55

from her practice.

2:56 – 2:57

And so in her retirement, she

2:57 – 3:00

moved back to her family's farm

3:00 – 3:02

and took over. And so in her

3:02 – 3:03

70s, she was farming

3:03 – 3:05

organically, like in 1996,

3:06 – 3:08

as an experimental,

3:09 – 3:09

creative,

3:10 – 3:11

pushing the edges of science,

3:11 – 3:12

thinking about the whole system

3:12 – 3:13

health.

3:14 – 3:15

So she had

3:15 – 3:16

converted her family's farm in

3:16 – 3:17

the late 90s to organic

3:17 – 3:18

production.

3:19 – 3:20

Met my parents at a farm

3:20 – 3:21

conference.

3:21 – 3:22

Sit down at lunch,

3:23 – 3:24

stand next to somebody and say,

3:24 – 3:25

who are you? What are you up to?

3:26 – 3:27

Hit off conversation.

3:28 – 3:30

Uh, and my, my dad had grown up

3:30 – 3:32

on a farm in Des Moines, Iowa,

3:33 – 3:34

big family, 10 kids, the farm

3:34 – 3:35

was lost in the seventies.

3:36 – 3:37

Parents said,

3:37 – 3:38

whatever you do, don't farm,

3:39 – 3:40

you know, stay away from it.

3:40 – 3:40

It's a tough way to live.

3:41 – 3:42

But there was sort of this

3:42 – 3:45

through line of, of being

3:45 – 3:46

connected to the land, wanting

3:46 – 3:47

to raise a family on the land.

3:47 – 3:50

And so my parents met me about

3:50 – 3:52

this conference and struck up a

3:52 – 3:53

conversation.

3:52 – 3:54

And she was in her eighties,

3:54 – 3:56

really young. to re -retire.

3:56 – 3:58

And everyone in the neighborhood

3:58 – 3:59

said, why don't I come by your

3:59 – 4:00

farm? And she said, that sounds

4:00 – 4:01

great.

4:01 – 4:02

It has to stay organic.

4:02 – 4:04

And nobody wanted to do that at

4:04 – 4:05

that time. So she was out

4:05 – 4:07

looking for somebody and met my

4:07 – 4:08

parents.

4:08 – 4:10

So we started,

4:10 – 4:12

we moved to the farm in the year

4:12 – 4:13

2000, 25 years ago,

4:14 – 4:15

very humble beginnings,

4:16 – 4:17

200 acres, borrowed money on a

4:17 – 4:18

4440.

4:19 – 4:21

and was really, really the early

4:21 – 4:22

days. And so that's the origin

4:22 – 4:23

story of our farm.

4:23 – 4:25

And I share that because getting

4:26 – 4:28

into farming is like the mafia.

4:28 – 4:30

You need to know somebody or

4:30 – 4:30

have a lot of money.

4:31 – 4:33

It's incredibly difficult.

4:33 – 4:35

And it's an issue that I'm

4:35 – 4:36

passionate about, thinking about

4:36 – 4:37

beginning farmers and

4:37 – 4:39

repopulating the land and

4:39 – 4:40

connecting communities to food.

4:41 – 4:43

And you just can't do it in

4:43 – 4:45

conventional commodity

4:45 – 4:46

agriculture.

4:46 – 4:47

And so that was the story of our

4:47 – 4:49

farm, is that how could you

4:49 – 4:50

support

4:51 – 4:52

a family.

4:51 – 4:53

And so we looked at the organic

4:53 – 4:55

niche. We're considering it for

4:56 – 4:57

environmental, ethical,

4:58 – 4:59

religious beliefs,

4:59 – 5:00

wanting to be stewards of the

5:00 – 5:01

land, but then also the

5:01 – 5:02

economics at that time.

5:04 – 5:05

A higher price for the

5:05 – 5:06

commodity,

5:06 – 5:08

niche markets, direct

5:08 – 5:09

relationships, all those things.

5:09 – 5:11

And so our family was able to

5:11 – 5:12

start

5:11 – 5:13

with basically nothing.

5:14 – 5:15

And over the last 25 years,

5:15 – 5:17

we've grown to about a 2 ,500

5:17 – 5:18

acre farm. So really parallel to

5:18 – 5:19

the growth of the industry.

5:21 – 5:22

Much of that growth has been

5:22 – 5:24

still in the grain world.

5:24 – 5:25

So we produce corn,

5:25 – 5:27

soybeans. Those often end up in

5:27 – 5:28

the feed grade market that end

5:28 – 5:30

up in feed for chickens and

5:30 – 5:31

livestock.

5:31 – 5:32

Some as food grade,

5:33 – 5:34

some distilling,

5:35 – 5:37

some food grade soybeans for end

5:37 – 5:38

users like

5:38 – 5:40

tofu, soy sauce,

5:40 – 5:41

wheat.

5:41 – 5:42

We have some canning crops

5:42 – 5:43

available to us.

5:43 – 5:44

So we do peas

5:44 – 5:47

and sweet corn and then edible

5:47 – 5:48

beans, black beans and kidney

5:48 – 5:49

beans, which are really fun.

5:50 – 5:50

And then some cover crops in

5:50 – 5:52

that rotation. But we're here on

5:52 – 5:53

the edge of the prairie.

5:53 – 5:56

So our ecosystem is really right

5:56 – 5:58

where the big woods hits prairie

5:58 – 5:59

pothole,

5:59 – 6:00

and it gets flat and wide.

6:00 – 6:01

And so both the landscape

6:01 – 6:03

changes and the type of farming

6:03 – 6:03

changes.

6:03 – 6:04

6:04 – 6:05

20 miles to the east of us, it's

6:05 – 6:06

rolling hills,

6:07 – 6:09

old dairies and silo harvesters

6:09 – 6:10

and 40 acre fields.

6:10 – 6:12

And 20 miles to the west is,

6:12 – 6:14

you know,

6:14 – 6:15

grain infrastructure that looks

6:15 – 6:17

like it could be a whole

6:17 – 6:18

cooperative and it's just one

6:18 – 6:19

individual farm and they're

6:20 – 6:21

20 ,000 acre farms.

6:21 – 6:23

So we're nestled kind of in this

6:23 – 6:24

geography,

6:24 – 6:27

cultural change, and we've

6:28 – 6:30

wiggled our way through it as

6:30 – 6:31

kind of a 1 .5 generation farm

6:31 – 6:32

with the organics.

6:32 – 6:34

So that financial stewardship

6:34 – 6:36

and really observing the

6:37 – 6:38

loss of the small family farm

6:38 – 6:39

and the

6:40 – 6:41

mega farms emerging over the

6:41 – 6:42

last 25 years.

6:47 – 6:48

So we've touched on the

6:48 – 6:50

financial management piece and

6:50 – 6:52

the need for kind of economic

6:52 – 6:54

independence in this very

6:54 – 6:55

challenged world.

6:54 – 6:56

You also touched on

6:56 – 6:59

the difficulty, the barriers to

6:59 – 7:01

entry, the difficulty of getting

7:01 – 7:02

in as something that you're

7:02 – 7:03

quite passionate about.

7:05 – 7:06

What's the intersection of those

7:06 – 7:07

two things and how are you

7:07 – 7:08

thinking about them in the

7:08 – 7:09

context of your operation today?

7:11 – 7:12

Well, we do a couple of things

7:12 – 7:13

on our farm. So we're trying to

7:13 – 7:15

create opportunities for other

7:15 – 7:16

farmers to pay that forward.

7:16 – 7:17

I think that's like a real

7:18 – 7:19

a real commitment that there was

7:19 – 7:22

a gift from Mabel to our family,

7:22 – 7:23

and we want to pay that forward.

7:23 – 7:25

So we've

7:25 – 7:27

dedicated ourselves to helping

7:27 – 7:29

other farmers get into it, and

7:29 – 7:30

especially folks making the

7:30 – 7:31

transition to organic.

7:31 – 7:33

So anytime somebody gives me a

7:33 – 7:35

call and says, hey, I'm thinking

7:35 – 7:35

about organics,

7:36 – 7:37

the door flies open.

7:38 – 7:39

We invite them in, start having

7:39 – 7:41

that conversation, talking about

7:41 – 7:43

strategies, transition,

7:44 – 7:45

market relationships,

7:48 – 7:49

7:49 – 7:51

father, Joe, has really mentored

7:51 – 7:53

dozens of folks over the years

7:53 – 7:56

and we've helped get folks

7:56 – 7:59

started and really try and

7:59 – 8:00

continue that opportunity

8:00 – 8:02

creation that was given to us

8:02 – 8:03

and pay that forward.

8:05 – 8:06

I think about grain farming

8:06 – 8:08

specifically as a more

8:08 – 8:09

sustainable way than the

8:10 – 8:11

the vegetable world.

8:11 – 8:13

A lot of my friends, other

8:13 – 8:16

idealistic millennials that were

8:16 – 8:17

out there trying to change the

8:17 – 8:18

world got into vegetable

8:18 – 8:19

farming.

8:19 – 8:20

What I see is a pretty

8:20 – 8:22

consistent pattern of folks

8:23 – 8:24

going to intern on a farm,

8:25 – 8:26

learning about farming.

8:27 – 8:29

Maybe they're 17, 18, 19,

8:29 – 8:31

into their early 20s.

8:31 – 8:32

They maybe figure out how to buy

8:32 – 8:34

20 acres of land in their early

8:34 – 8:36

20s. They hit 35.

8:36 – 8:37

They've got two kids.

8:37 – 8:38

their bodies start to break

8:38 – 8:40

down, they realize,

8:40 – 8:41

uh -oh, this isn't working out.

8:41 – 8:42

The economics aren't working

8:42 – 8:44

out. So at least here in the

8:44 – 8:46

Midwest context, I'm always

8:46 – 8:47

encouraging folks that are

8:47 – 8:48

beginning farmers or first

8:48 – 8:49

-generation farmers to think

8:49 – 8:50

about

8:50 – 8:53

pushing beyond just the quick

8:53 – 8:55

natural solution of being a CSA

8:55 – 8:56

vegetable farmer.

8:56 – 8:57

I think that's a really

8:57 – 8:58

important part of our food

8:58 – 8:59

system, but it's still

8:59 – 9:00

undervalued.

9:00 – 9:02

And so grain farming is a huge

9:02 – 9:03

opportunity for beginning

9:03 – 9:05

farmers because there are

9:05 – 9:07

systems in place, there's scale.

9:07 – 9:08

To a certain amount, there's

9:08 – 9:09

mechanization

9:09 – 9:11

helps protect your body over

9:11 – 9:12

time.

9:14 – 9:16

And economics are still a little

9:16 – 9:17

bit more robust and a little bit

9:17 – 9:18

more stable.

9:18 – 9:20

So at least start with grain and

9:20 – 9:21

think about maybe adding

9:21 – 9:22

livestock for diversification.

9:23 – 9:24

If you can complete the whole

9:24 – 9:27

cycle with vegetables or direct

9:27 – 9:29

-to -consumer stuff, that's kind

9:29 – 9:31

of the sweet spot, I think, for

9:31 – 9:33

a really sustainable economic

9:33 – 9:34

farm.

9:34 – 9:37

So when you think about grain

9:37 – 9:39

farming, of course, even organic

9:39 – 9:41

farming, there is still a

9:41 – 9:43

substantial barrier to entry

9:43 – 9:44

relative to a fruit and

9:44 – 9:46

vegetable farm, relative to the

9:46 – 9:48

size that is needed and the

9:48 – 9:49

acreage that is needed.

9:49 – 9:51

So how do you solve that barrier

9:51 – 9:52

to entry?

9:53 – 9:56

Good question. It's not easy.

9:57 – 9:58

I certainly don't want to

9:59 – 10:00

suggest that it's just,

10:00 – 10:02

you know, go buy a tractor and

10:02 – 10:02

get going.

10:05 – 10:06

We have an opportunity because

10:06 – 10:07

there's a huge generational

10:07 – 10:08

shift, right? I think half of

10:08 – 10:09

the land is going to change

10:09 – 10:11

hands over the next decade.

10:12 – 10:15

And so that changing of hands is

10:15 – 10:16

the opportunity, I think,

10:16 – 10:18

because just

10:19 – 10:20

going out and being able to buy

10:20 – 10:21

a farm is

10:21 – 10:22

darn near impossible,

10:23 – 10:24

given how expensive land is.

10:24 – 10:26

In our context here, Minnesota

10:26 – 10:28

land is $11 ,000 or $12 ,000 an

10:28 – 10:29

acre.

10:30 – 10:31

Tractors can be hundreds of

10:31 – 10:32

thousands of dollars, combines.

10:32 – 10:34

So I think we're

10:34 – 10:35

a beginning farmer can think

10:35 – 10:36

about,

10:36 – 10:38

partnering where they can maybe

10:38 – 10:39

offer their labor, their energy,

10:40 – 10:42

their human capital over their

10:42 – 10:43

financial capital in partnership

10:43 – 10:44

with farmers,

10:45 – 10:46

retiring folks, folks who are

10:46 – 10:47

aging out.

10:50 – 10:51

One of the things that we did

10:51 – 10:53

here in Minnesota is worked on a

10:53 – 10:54

beginning farmer tax credit.

10:55 – 10:56

So in 2017,

10:57 – 10:58

I worked with the National Young

10:58 – 11:00

Farmers Association and we've

11:00 – 11:01

worked on a beginning farmer tax

11:01 – 11:03

credit that incentivizes the

11:03 – 11:06

sale or lease of agricultural

11:06 – 11:07

assets. That's land and

11:07 – 11:08

equipment.

11:08 – 11:09

to beginning farmers.

11:10 – 11:12

So there's some kind of policy,

11:12 – 11:14

financial mechanisms you can try

11:14 – 11:14

and do.

11:17 – 11:19

One of my marketing techniques

11:19 – 11:23

is to find folks who are

11:23 – 11:25

retiring and mail them boxes of

11:25 – 11:25

our product.

11:25 – 11:28

So as an example, we have black

11:28 – 11:29

beans.

11:29 – 11:31

I literally mail people cans of

11:31 – 11:32

our black beans and say, hey,

11:32 – 11:33

I'm in the neighborhood.

11:33 – 11:34

We're a beginning farmer.

11:34 – 11:35

Here's our store.

11:35 – 11:37

We raise food and it's good and

11:37 – 11:38

it tastes good.

11:38 – 11:39

And this is what it looks like.

11:39 – 11:40

And if you're interested in

11:40 – 11:41

something different, I'd love to

11:41 – 11:42

talk.

11:42 – 11:43

Um,

11:43 – 11:44

so it still requires kind of

11:44 – 11:46

being scrappy and being creative

11:46 – 11:47

to get your way into them.

11:49 – 11:51

What you're describing is

11:52 – 11:54

particularly that your, your

11:54 – 11:55

last, um,

11:55 – 11:56

thing that you described or

11:56 – 11:57

reaching out to people, you, you

11:57 – 11:58

were already established.

11:58 – 12:01

You already have a foothold and,

12:03 – 12:04

I think there are, there are

12:04 – 12:07

many young people who want to

12:07 – 12:07

get in,

12:08 – 12:10

who want to get started, who've

12:10 – 12:12

become quite cynical about the

12:12 – 12:13

older generation.

12:14 – 12:15

Um, you had an exceptional

12:15 – 12:17

opportunity in connecting with

12:17 – 12:19

someone who cared more deeply

12:19 – 12:20

about mission than they cared

12:20 – 12:21

about money.

12:22 – 12:23

And,

12:23 – 12:25

um, the reality is many,

12:26 – 12:27

uh, and this is

12:27 – 12:28

not true for everyone, but

12:28 – 12:30

there's this cynical perspective

12:30 – 12:31

by the younger generation that

12:31 – 12:32

the older generation is just

12:32 – 12:34

going to cash out, sell out to

12:34 – 12:35

investors, get as much money as

12:35 – 12:36

they can that they.

12:36 – 12:37

They care more about the money

12:37 – 12:38

than they care about the next

12:38 – 12:39

generation.

12:40 – 12:41

So how do we navigate that

12:41 – 12:42

landscape?

12:45 – 12:46

I kind of partially agree with

12:46 – 12:47

the cynicism.

12:47 – 12:48

You know, it's not it's not just

12:48 – 12:49

made up.

12:50 – 12:50

12:52 – 12:53

operate within a capitalist

12:53 – 12:55

society that drives the

12:56 – 12:57

most return on investment, the

12:57 – 12:58

highest amount of money

12:58 – 12:59

possible. It's it's very

12:59 – 13:00

tempting.

13:00 – 13:02

So I can be empathetic with the

13:02 – 13:03

folks that make that choice

13:03 – 13:05

because that

13:06 – 13:07

that's very, very powerful.

13:10 – 13:11

And frankly, it's going to be

13:11 – 13:13

hard to beat. So I think you do

13:13 – 13:14

have to just place yourself in

13:14 – 13:15

spaces where people are mission

13:15 – 13:16

aligned.

13:17 – 13:19

The reason that meeting happened

13:19 – 13:21

between my parents and Mabel was

13:21 – 13:22

they were at a conference with

13:23 – 13:24

like -minded folks thinking

13:24 – 13:24

about organics.

13:25 – 13:26

So if it's the Acres Conference,

13:27 – 13:29

if it's Marble Seed, if it's a

13:29 – 13:30

community gathering, around

13:30 – 13:32

sustainable food, if it's going

13:32 – 13:33

to your farmer's market and

13:33 – 13:35

introducing yourself to the

13:35 – 13:36

folks that are at the farmer's

13:36 – 13:37

market,

13:38 – 13:40

I think placing yourself in

13:40 – 13:41

spaces where you maybe find

13:41 – 13:43

those like -minded folks is

13:43 – 13:44

probably the best way to do

13:44 – 13:45

that.

13:46 – 13:47

The cold calling that I

13:47 – 13:48

described is really difficult.

13:49 – 13:51

That's almost a used car

13:51 – 13:52

salesman gig.

13:52 – 13:55

That's an uphill battle.

13:55 – 13:56

I don't know if I would describe

13:56 – 13:57

it that way.

14:00 – 14:01

I've sent lots of cans of black

14:01 – 14:04

beans and I've only had

14:04 – 14:05

one or two opportunities come

14:05 – 14:06

out of that tactic.

14:06 – 14:07

It's not like a

14:07 – 14:11

perfect tool.

14:12 – 14:14

But yeah, your point's well

14:14 – 14:16

taken. It is incredibly

14:16 – 14:17

difficult, and I think it is OK

14:17 – 14:18

to be

14:18 – 14:19

realistic about that.

14:19 – 14:21

And so creating

14:22 – 14:24

opportunity, creating luck,

14:24 – 14:25

creating those connections,

14:25 – 14:27

you know, set yourself up in an

14:27 – 14:28

environment where that could

14:28 – 14:29

happen.

14:31 – 14:33

Yeah, we could have we could

14:33 – 14:35

have a lot more conversations

14:35 – 14:36

around that topic.

14:36 – 14:37

But I want to dig into

14:37 – 14:39

the

14:39 – 14:40

way that you set your farm up

14:40 – 14:42

for financial success, both

14:42 – 14:44

historically and now.

14:44 – 14:46

I think a critical part of this

14:47 – 14:47

story

14:48 – 14:49

particularly in today's

14:49 – 14:50

environment, although I would

14:50 – 14:51

argue it's been true for the

14:51 – 14:52

last couple of decades,

14:52 – 14:54

is the need to decommoditize

14:54 – 14:58

yourself by any one of dozens of

14:58 – 14:59

possible different mechanisms,

14:59 – 15:00

which I think you have done.

15:01 – 15:02

So let's

15:03 – 15:04

talk a bit about your

15:05 – 15:06

thought process.

15:06 – 15:07

How do you think about your

15:07 – 15:10

farm's economic viability and

15:10 – 15:11

sustainability?

15:11 – 15:12

You've also grown.

15:12 – 15:13

You've grown 10x over the last

15:13 – 15:14

couple of decades.

15:14 – 15:15

How were you able to accomplish

15:15 – 15:16

that?

15:22 – 15:24

There's probably two or three

15:24 – 15:25

key components.

15:25 – 15:27

I'd say one is the partnerships

15:27 – 15:28

in your team.

15:28 – 15:29

The second is that

15:29 – 15:30

decommodification.

15:35 – 15:36

Why don't we tackle those first

15:36 – 15:37

two things, and then I'll come

15:37 – 15:38

back to the third thing I'm

15:38 – 15:39

thinking about.

15:38 – 15:39

But the first thing is a

15:39 – 15:40

partnership and team.

15:41 – 15:42

And both, it's like your

15:42 – 15:43

internal team.

15:43 – 15:45

So that scale requires having

15:46 – 15:47

folks who are on the farm every

15:47 – 15:48

day who are bought in, who

15:48 – 15:49

believe in that mission, who are

15:49 – 15:51

willing to work, who are

15:51 – 15:52

bringing their whole selves.

15:54 – 15:55

And so we have a really, really

15:55 – 15:56

strong team on the farm.

15:56 – 15:58

I farm with my father, Joe.

15:58 – 15:59

We have one or two full -time

15:59 – 16:00

folks. We have some retired

16:00 – 16:02

farmers who help us out who are

16:02 – 16:03

excited about the organic

16:04 – 16:05

thing. It reminds them of how

16:05 – 16:06

they grew up.

16:08 – 16:10

And when we meet, we talk about

16:11 – 16:12

why we do what we do,

16:13 – 16:15

our care of stewardship of

16:15 – 16:16

the land.

16:16 – 16:19

our commitment to proving that

16:19 – 16:20

sustainable agriculture is

16:20 – 16:21

competitive.

16:21 – 16:23

We play the underdog card a

16:23 – 16:25

little bit. We motivate everyone

16:25 – 16:26

around that,

16:27 – 16:28

proving that we can do it.

16:30 – 16:32

And then it's the other partners

16:32 – 16:34

that we have. So it's financial

16:34 – 16:35

partners.

16:35 – 16:37

We bank with a group called MAD

16:37 – 16:38

Agriculture,

16:38 – 16:40

which is an organization that's

16:40 – 16:41

committed to organics and

16:41 – 16:42

regenerative agriculture.

16:42 – 16:44

So they share in that vision and

16:44 – 16:45

mission. So some of their

16:45 – 16:46

financial tools are more

16:47 – 16:49

built and designed for organic

16:49 – 16:50

and regenerative farms, as

16:50 – 16:51

opposed to just a conventional

16:51 – 16:53

commercial agricultural lender

16:53 – 16:55

that maybe doesn't have the

16:55 – 16:56

context, doesn't understand the

16:56 – 16:57

cash flows.

16:58 – 17:01

It's having peer

17:01 – 17:02

farmers.

17:02 – 17:05

So on my phone are the speed

17:05 – 17:06

dial numbers of farmers who are

17:06 – 17:08

like me that I can talk to and

17:08 – 17:09

work through problems,

17:09 – 17:10

questions.

17:11 – 17:12

I'd

17:13 – 17:14

like to say having a good

17:14 – 17:15

agronomist, but that's really

17:15 – 17:16

difficult.

17:16 – 17:17

So I'm excited to talk about it.

17:17 – 17:19

Let me translate that.

17:32 – 17:33

There's a myth that a farmer has

17:33 – 17:34

to be everything.

17:34 – 17:36

And we know that that's not true

17:36 – 17:37

with the plants that we raise,

17:37 – 17:39

that the crops we're raising are

17:39 – 17:40

whole systems. And we have to

17:40 – 17:41

treat ourselves and our

17:41 – 17:42

businesses as whole systems.

17:42 – 17:44

And we have to bring in all the

17:44 – 17:45

right elements and have them

17:45 – 17:46

complement and interact with

17:46 – 17:47

each other. So that's how I

17:47 – 17:48

think about the partnership that

17:48 – 17:49

allows that

17:50 – 17:52

growth, that financial

17:52 – 17:53

stability. Because we have

17:53 – 17:55

hiccups, we have droughts, we

17:55 – 17:56

have crop failures, we have

17:57 – 17:59

grain buyers that,

17:59 – 18:00

you know,

18:00 – 18:01

go broke.

18:02 – 18:04

We are not without problems,

18:04 – 18:05

challenges,

18:05 – 18:06

and things going sideways.

18:07 – 18:09

I think one of the things that

18:09 – 18:11

you said that stood

18:12 – 18:14

out to me was your discussion of

18:14 – 18:16

the team of people that you have

18:16 – 18:16

around you.

18:17 – 18:18

It's

18:18 – 18:19

one of the things that I see as

18:19 – 18:21

being so foundational.

18:21 – 18:23

Many farms have this to varying

18:23 – 18:24

degrees,

18:24 – 18:25

but

18:27 – 18:27

18:28 – 18:29

in understanding plant

18:29 – 18:30

communities and microbial

18:30 – 18:32

communities we're developing

18:32 – 18:33

this emerging understanding of

18:33 – 18:35

quorum sensing and how how the

18:35 – 18:37

entire system changes the

18:38 – 18:40

behavior of the entire system

18:40 – 18:41

changes when you have enough

18:41 – 18:44

diversity of perspectives to

18:44 – 18:46

establish quorum sensing and

18:46 – 18:47

it's kind of the same way in

18:47 – 18:49

human organizations or human

18:49 – 18:50

communities when you have enough

18:50 – 18:51

diversity of perspectives like

18:51 – 18:53

you need enough different points

18:53 – 18:55

of view if you only have

18:56 – 18:58

five or ten people and they all

18:58 – 18:59

share the identical point of

18:59 – 19:02

view, then you only have a much

19:02 – 19:03

more limited perspective.

19:04 – 19:05

You have much less resilience.

19:07 – 19:07

Yeah.

19:08 – 19:08

It's

19:09 – 19:10

hard to do it too.

19:10 – 19:12

I think that it's not an easy

19:12 – 19:13

thing.

19:15 – 19:17

And it's long, slow work, just

19:17 – 19:18

like stewardship or soil health.

19:19 – 19:21

It's long, slow work to build

19:21 – 19:22

that. And so I absolutely

19:22 – 19:24

personally benefit from the work

19:24 – 19:26

that my parents have done before

19:26 – 19:27

me. And I feel like the work I'm

19:27 – 19:28

doing today is for the next

19:28 – 19:30

generation or for the

19:31 – 19:32

next farmer. And if you just see

19:32 – 19:33

yourself as part of this larger

19:33 – 19:34

arc,

19:34 – 19:35

this larger ecosystem,

19:36 – 19:37

it brings joy to the work, too,

19:37 – 19:38

frankly.

19:39 – 19:40

Sometimes, this is just hard.

19:40 – 19:41

There's no way around it.

19:41 – 19:42

It's just hard.

19:42 – 19:42

It's stressful.

19:44 – 19:46

For me, at least, it's the

19:47 – 19:47

human element.

19:47 – 19:48

The community element is really

19:49 – 19:50

grounding

19:51 – 19:52

and life -giving.

19:53 – 19:54

Yeah.

19:54 – 19:54

19:55 – 19:56

want to not lose sight of the

19:56 – 19:57

third thing that you had in the

19:57 – 19:57

back of your mind.

19:57 – 19:58

Oh, yeah, yeah, yeah.

19:58 – 19:59

So,

19:59 – 20:01

I talked about partnership

20:02 – 20:03

decommodification really

20:03 – 20:04

briefly.

20:04 – 20:06

So, I would say we're still

20:06 – 20:07

probably

20:08 – 20:10

60 to 70 % of our crops are what

20:10 – 20:11

you might consider an organic

20:11 – 20:12

commodity.

20:12 – 20:13

You know, they go to a broker

20:13 – 20:14

that we don't know personally

20:14 – 20:15

that gets

20:16 – 20:17

shipped on a rail car to the

20:17 – 20:18

East Coast or West Coast and

20:18 – 20:19

ends up...

20:20 – 20:20

We're slowly...

20:20 – 20:21

There are degrees of

20:21 – 20:23

decommoditization, obviously,

20:23 – 20:24

but I would argue that the fact

20:24 – 20:25

that you are growing a crop

20:25 – 20:26

that's organically certified

20:26 – 20:28

already to some degree has

20:28 – 20:29

removed you from the mainstream

20:29 – 20:30

commodity marketplace.

20:31 – 20:32

Yeah. Yeah, I would say so.

20:32 – 20:34

Certainly we're...

20:34 – 20:35

Our corn

20:35 – 20:37

is worth twice as much as a...

20:37 – 20:39

kernel corn sold on the Chicago

20:39 – 20:40

border trade.

20:40 – 20:41

And it has a lot more

20:42 – 20:44

nutritional value and love and

20:44 – 20:45

effort put into it.

20:45 – 20:46

So I think that's a fair price.

20:48 – 20:49

But yeah,

20:50 – 20:52

the third component,

20:52 – 20:53

I'd say,

20:54 – 20:56

for like, how do you manage the

20:56 – 20:57

scale? Or how do you

20:57 – 20:58

de -risk things?

20:59 – 21:00

Is this

21:01 – 21:04

concept of doing the right thing

21:04 – 21:07

at the right time and in the

21:07 – 21:08

right way.

21:08 – 21:09

And what I want to talk about

21:09 – 21:11

here a project that I've been

21:11 – 21:12

working on for the last three

21:12 – 21:13

years called Farm Flow.

21:14 – 21:14

And I think,

21:15 – 21:16

John, you'll share this insight,

21:17 – 21:19

which is like alternative

21:19 – 21:20

systems require intensive

21:20 – 21:22

management. It's not a passive

21:22 – 21:23

endeavor.

21:23 – 21:24

And I think that's often what

21:24 – 21:25

people think organic

21:25 – 21:27

or regenerative or permaculture

21:27 – 21:28

or something. It's sort of like,

21:28 – 21:29

well, you just let

21:30 – 21:32

nature do its thing and it's

21:32 – 21:33

going to work its way out.

21:33 – 21:34

And it's not that at all.

21:34 – 21:36

nature is absolutely doing its

21:36 – 21:38

thing, but you're an active part

21:38 – 21:39

of it.

21:40 – 21:41

Actually, I'd like to touch on

21:41 – 21:43

that in just just a moment for

21:43 – 21:44

just from a very hot, extremely

21:44 – 21:46

high level philosophical or

21:46 – 21:47

principles perspective.

21:49 – 21:50

21:51 – 21:53

I've been learning more about

21:53 – 21:54

quorum sensing and how

21:54 – 21:55

ecosystems function,

21:56 – 21:58

I think there is an it's

21:58 – 21:59

possible to

22:00 – 22:02

develop an ecosystem and develop

22:02 – 22:04

a landscape where nature just

22:04 – 22:05

does its thing.

22:05 – 22:08

And you there there is much less

22:08 – 22:09

management involved.

22:09 – 22:09

I think

22:09 – 22:11

One of the early inspirations of

22:11 – 22:13

this outside of the ranching and

22:13 – 22:15

grazing world was Masanobu

22:15 – 22:16

Fukuoka and his one straw

22:16 – 22:18

revolution in natural farming

22:18 – 22:19

and some of the work that he was

22:19 – 22:20

doing.

22:20 – 22:21

But it's becoming increasingly

22:21 – 22:23

clear to me as I understand

22:23 – 22:26

and look at different examples

22:26 – 22:27

of this, that in order for that

22:27 – 22:29

principle to

22:30 – 22:31

even possibly hold true

22:32 – 22:34

requires an abundance of

22:34 – 22:35

different species and requires

22:35 – 22:36

an abundance of diversity.

22:36 – 22:38

That doesn't happen in organic

22:38 – 22:39

grain cropping system.

22:39 – 22:40

It doesn't happen in most

22:40 – 22:41

farming systems.

22:42 – 22:44

So I think that principle can be

22:44 – 22:47

true, but isn't true in most

22:48 – 22:50

grain cropping environments, or

22:50 – 22:51

even all of them.

22:51 – 22:52

Yeah. Yeah.

22:53 – 22:54

And there's this transition

22:54 – 22:55

period from what you're

22:55 – 22:56

describing,

22:56 – 22:58

this quorum sensing that like

22:58 – 22:59

requires active participation,

22:59 – 23:01

active management, active.

23:02 – 23:03

Yeah.

23:03 – 23:05

So management entails

23:05 – 23:06

absolutely.

23:06 – 23:08

Most, most grain cropping to do

23:08 – 23:09

it really well,

23:10 – 23:11

to do regenerative management,

23:11 – 23:12

organic management really well

23:12 – 23:14

requires much more knowledge.

23:14 – 23:15

It's much more knowledge

23:15 – 23:16

intensive and much more

23:16 – 23:16

management intensive.

23:17 – 23:18

You've, you've taken away all

23:18 – 23:18

the easy buttons.

23:20 – 23:21

Yep.

23:21 – 23:22

So three

23:23 – 23:24

and a half years ago I had a

23:24 – 23:25

friend who's a data scientist

23:25 – 23:26

and I said, Hey,

23:27 – 23:27

I have a bunch of information

23:27 – 23:29

about my farm. And I'd like to

23:29 – 23:30

think about the question,

23:31 – 23:32

This is a natural farmer

23:32 – 23:33

question.

23:33 – 23:34

How could we raise more yields?

23:34 – 23:35

How could we raise a bigger

23:35 – 23:36

crop?

23:36 – 23:38

And what could we do?

23:38 – 23:39

How could we figure that out?

23:40 – 23:41

And so he was willing to help me

23:41 – 23:42

on this project.

23:42 – 23:44

And so we put together a data

23:44 – 23:44

set.

23:44 – 23:46

And this isn't a university

23:46 – 23:47

-level research with randomized

23:47 – 23:50

trials. It's just our own multi

23:50 – 23:50

-year,

23:51 – 23:52

large -scale farm.

23:52 – 23:52

And we looked at

23:53 – 23:54

planting date, variety,

23:55 – 23:55

crop fertility,

23:56 – 23:57

soil suitability,

23:57 – 23:59

a whole bunch of different

23:59 – 24:00

factors on our farm, including

24:00 – 24:02

we looked at the number of

24:02 – 24:03

passes.

24:03 – 24:04

So the number of times you were

24:04 – 24:05

in a field doing some kind of

24:05 – 24:06

operation or scouting or

24:06 – 24:08

intervention or cover crop or

24:08 – 24:09

whatever it might be, and the

24:09 – 24:10

frequency.

24:10 – 24:13

So how often were those passes

24:13 – 24:13

occurring?

24:13 – 24:15

We ran this multivariable

24:15 – 24:16

regression, which basically

24:16 – 24:18

helps us understand like which

24:18 – 24:19

of these variables has the

24:19 – 24:20

largest impact on that

24:21 – 24:22

yield.

24:22 – 24:23

And when we ran that regression,

24:23 – 24:25

the strongest variable was the

24:25 – 24:26

frequency of passes.

24:27 – 24:28

So if we were in a field every

24:29 – 24:31

four days versus every seven

24:31 – 24:32

days, there was a five bushel

24:33 – 24:34

increase in corn yield.

24:35 – 24:38

So the insight was,

24:39 – 24:40

the more often you're in your

24:40 – 24:41

field, the better your crops.

24:42 – 24:43

And you're going to go, yeah,

24:43 – 24:44

duh.

24:44 – 24:45

You talk to any good gardener

24:45 – 24:46

and they say, yeah, if you weed

24:46 – 24:47

your garden,

24:47 – 24:48

the more often you're weeding,

24:49 – 24:49

the better your tomatoes.

24:50 – 24:52

But that insight really made

24:52 – 24:54

clear to us the

24:55 – 24:58

active intentional management of

24:58 – 24:58

our farms

24:59 – 25:00

drove outcrops.

25:00 – 25:01

that we wanted, whether it was

25:01 – 25:02

weed control, whether it was

25:02 – 25:03

yield, whether it was like the

25:03 – 25:04

success of a cover crop,

25:04 – 25:05

whatever it was.

25:06 – 25:08

And despite being good farmers

25:08 – 25:10

and having 25 years of

25:10 – 25:11

experience, we weren't always

25:11 – 25:11

doing that ourselves.

25:12 – 25:13

We kind of knew this insight.

25:14 – 25:15

The regression proved something

25:15 – 25:16

we already intuitively

25:16 – 25:17

understood,

25:18 – 25:18

but we had a gap.

25:20 – 25:22

And so we came up with a

25:22 – 25:23

solution for that gap,

25:23 – 25:25

which is the farm flow concept.

25:25 – 25:26

I can share more about that.

25:27 – 25:28

Yeah,

25:29 – 25:30

I have a

25:31 – 25:32

couple of flags that ran up the

25:32 – 25:33

poll for me here.

25:34 – 25:37

So one of them is this...

25:37 – 25:38

When we look at

25:39 – 25:41

plant nutrition management from

25:41 – 25:42

a physiological development

25:42 – 25:44

perspective and

25:44 – 25:46

the phenological growth stages,

25:47 – 25:48

it's become

25:48 – 25:52

extremely clear that there are

25:52 – 25:54

windows of time where

25:54 – 25:55

we can do this.

25:55 – 25:57

Uh, if you get on the necessary

25:57 – 25:59

application, let's say I'm just

25:59 – 26:00

making things up here, but let's

26:00 – 26:01

say you get on a critical,

26:02 – 26:04

um, side dress application on

26:04 – 26:06

corn at V5. And if you miss

26:06 – 26:08

that, if you have like a three

26:08 – 26:09

to five day window, and if you

26:09 – 26:11

miss that window and you're a

26:11 – 26:12

week late,

26:12 – 26:14

then your yield response is down

26:14 – 26:15

disproportionately.

26:15 – 26:17

Like you can, that can cost you

26:17 – 26:19

10 bushels in yield and there

26:19 – 26:20

are all these various windows.

26:20 – 26:22

So on one hand, when you talk

26:22 – 26:23

about,

26:23 – 26:25

I understand the, the critical

26:25 – 26:26

need.

26:26 – 26:28

to get certain management

26:28 – 26:31

activities done within very well

26:31 – 26:32

-defined windows.

26:32 – 26:33

I understand how important that

26:33 – 26:34

is.

26:34 – 26:35

But that's not exactly what you

26:35 – 26:37

said. You said the

26:38 – 26:40

frequency of and the number of

26:40 – 26:42

times that you're in the field.

26:42 – 26:43

And so

26:43 – 26:46

if you take that all the way to

26:46 – 26:47

the extreme, you could say,

26:47 – 26:48

well, if you're in the field

26:49 – 26:51

every three days then for the

26:51 – 26:52

entire season, that should

26:52 – 26:53

produce the highest yield,

26:53 – 26:53

right?

26:54 – 26:55

But I don't think that's the

26:55 – 26:56

case. what you're saying.

26:56 – 26:58

No, it isn't. And, uh, you know,

26:58 – 27:00

that's a limit of a farmer, a

27:00 – 27:01

farmer run, uh,

27:02 – 27:03

data set and experiment, you

27:03 – 27:04

know, that there's, there's, uh,

27:05 – 27:06

an insight that can be clean,

27:06 – 27:07

but if you take it too far, it

27:07 – 27:09

falls apart, you know, or it's

27:09 – 27:10

limited in its utility.

27:12 – 27:13

So,

27:13 – 27:14

um,

27:16 – 27:18

From this insight, we created a

27:18 – 27:20

$300 idea.

27:20 – 27:21

I think it's worth a million

27:21 – 27:22

dollars.

27:23 – 27:25

We can give it to everybody for

27:25 – 27:25

free.

27:26 – 27:28

We basically took this insight

27:28 – 27:29

of like, okay, we need to

27:29 – 27:30

visualize the work that's being

27:30 – 27:33

done on the farm and catch those

27:33 – 27:36

gaps to get that side dress at

27:36 – 27:37

that right time.

27:38 – 27:40

We took a 4x8 whiteboard

27:40 – 27:42

and we took a tiny little tape

27:42 – 27:44

and we made a grid,

27:45 – 27:46

4x8 grid.

27:49 – 27:50

listed our fields, and from left

27:50 – 27:51

to right,

27:51 – 27:52

counted out days.

27:52 – 27:55

So individual squares equaled an

27:55 – 27:56

individual day on an individual

27:56 – 27:57

farm.

27:58 – 27:59

And then we just simply used

27:59 – 27:59

magnets.

27:59 – 28:01

And we placed a magnet with

28:01 – 28:03

a color code. So for example,

28:03 – 28:04

planting was green,

28:05 – 28:06

time weeding was blue,

28:06 – 28:07

manure application,

28:07 – 28:09

brown, so on and so forth.

28:09 – 28:11

And what we were able to do was

28:11 – 28:12

literally just chart out the

28:12 – 28:13

work that was happening on the

28:13 – 28:14

farm by placing magnets.

28:14 – 28:16

So you could walk into the

28:16 – 28:17

gathering shop,

28:17 – 28:19

and look at this giant

28:19 – 28:20

whiteboard, and without even

28:20 – 28:21

reading any details, you could

28:21 – 28:22

visualize,

28:23 – 28:24

here are these patterns of work

28:24 – 28:25

that's occurring from seeing the

28:25 – 28:28

colors,

28:29 – 28:30

and then the gaps in the colors.

28:31 – 28:31

And so really quickly,

28:32 – 28:34

that team, retired farmers, part

28:34 – 28:35

-time seasonal, full -time

28:35 – 28:37

folks, all could quickly

28:37 – 28:38

understand what was happening on

28:38 – 28:39

the farm,

28:39 – 28:40

what was going on,

28:41 – 28:42

what was getting behind.

28:43 – 28:44

And all of a sudden,

28:45 – 28:46

what changed was we

28:47 – 28:49

were able to respond, maybe a

28:49 – 28:50

three -inch range,

28:50 – 28:50

occurred and

28:51 – 28:52

projects were delayed, or a

28:52 – 28:55

piece of equipment broke, or a

28:55 – 28:57

farm that was maybe further away

28:57 – 28:58

or had a grumpy neighbor that we

28:58 – 29:00

had a subconscious bias, no one

29:00 – 29:00

wanted to go work at.

29:02 – 29:04

We could identify those things

29:04 – 29:05

and respond to them.

29:05 – 29:07

And so people had a sense of

29:07 – 29:09

their role, their participation,

29:09 – 29:10

their impact in the whole

29:10 – 29:10

system.

29:12 – 29:13

Folks that didn't, were new to

29:13 – 29:15

organics could start to

29:15 – 29:16

understand the sequence of these

29:16 – 29:18

interventions, these passes and

29:18 – 29:19

how they complemented each other

29:19 – 29:20

and how they built on each

29:20 – 29:21

other.

29:22 – 29:23

And as a result of just this

29:23 – 29:24

29:24 – 29:26

$300 whiteboard with magnets up

29:26 – 29:27

on it,

29:27 – 29:28

we changed the way that we were

29:28 – 29:29

thinking about the farm.

29:30 – 29:31

Everyone had a high level buy

29:31 – 29:33

-in, we were executing much

29:33 – 29:34

more quickly, we're catching

29:34 – 29:36

those gaps, that management gap,

29:37 – 29:40

and it changed the flow of

29:40 – 29:41

our farm. And so that was the

29:41 – 29:42

farm flow concept.

29:46 – 29:47

I want to talk about

29:47 – 29:48

the results that you

29:49 – 29:50

observed and what changed.

29:51 – 29:52

But before we go there,

29:53 – 29:55

let's dig into a bit more on

29:55 – 29:56

what

29:56 – 29:58

this chart actually looked like

29:58 – 29:59

on the wall. So I'm imagining

29:59 – 30:01

that over the course of a

30:01 – 30:02

season,

30:02 – 30:05

you have some gaps that are

30:05 – 30:05

three

30:06 – 30:07

days apart and you have some

30:07 – 30:09

that are 30 days apart.

30:10 – 30:11

30:11 – 30:13

you deliberately try to bring

30:13 – 30:15

things closer together and have

30:15 – 30:16

greater frequency?

30:16 – 30:18

Or are you just trying to hit

30:18 – 30:19

those optimal time windows?

30:23 – 30:24

It's seasonally appropriate.

30:25 – 30:26

So, you know, the activities

30:26 – 30:27

that are happening in August and

30:27 – 30:29

September are much more spread

30:29 – 30:31

out than they might be in May

30:31 – 30:32

and June on our farm in a green

30:32 – 30:33

system.

30:33 – 30:34

So in May and June, I might

30:34 – 30:35

expect something to be every

30:35 – 30:37

three to five days.

30:37 – 30:38

And by the time we get to

30:38 – 30:39

September, it might be once

30:39 – 30:40

every 30

30:40 – 30:41

days. You know, we might be

30:42 – 30:44

applying fertility and seeding a

30:44 – 30:45

cover crop, and that might be

30:45 – 30:46

only activity within those 60

30:46 – 30:47

days.

30:47 – 30:48

So,

30:48 – 30:49

30:49 – 30:50

can, I don't know if you're,

30:51 – 30:52

it won't be helpful to the

30:52 – 30:53

listeners, but I could share an

30:53 – 30:55

image of it for your benefit,

30:55 – 30:56

John.

30:56 – 30:58

Just describe it visually, I

30:58 – 30:59

think.

30:59 – 31:02

Yeah, so there's a limit to a

31:02 – 31:04

four -by -eight whiteboard

31:04 – 31:06

because you can't fit 365 tiny

31:06 – 31:08

little squares on a four -by

31:08 – 31:09

-eight. I think that means a

31:09 – 31:10

second four -by -eight

31:09 – 31:11

whiteboard. Yeah, we just keep

31:11 – 31:13

stacking together or, you know,

31:13 – 31:15

take a photo, you wipe it off.

31:15 – 31:16

I also have a two -year -old, so

31:16 – 31:18

we have to be careful about how

31:18 – 31:19

high those magnets are because

31:19 – 31:21

she can come in and move them

31:21 – 31:22

around pretty quickly and the

31:22 – 31:23

whole plan's off.

31:23 – 31:24

So,

31:24 – 31:25

yeah,

31:27 – 31:29

the tool itself really became

31:31 – 31:32

a gathering point or a

31:32 – 31:33

conversation for you were

31:33 – 31:35

talking about the diversity of

31:35 – 31:36

opinions.

31:36 – 31:37

And so

31:37 – 31:39

we all knew what the reality

31:39 – 31:40

was, but then we could talk

31:40 – 31:42

about in the morning, like, hey,

31:42 – 31:43

what's going on?

31:43 – 31:44

What are people seeing?

31:44 – 31:45

What do we think we need to do?

31:45 – 31:48

And it gave us a reference point

31:48 – 31:48

that we could start to talk

31:48 – 31:50

about as a team of planning out,

31:50 – 31:51

because before really everything

31:51 – 31:52

lives in your head.

31:53 – 31:54

Or maybe if you're hyper

31:54 – 31:55

organized, maybe in a

31:55 – 31:57

spreadsheet or on a

31:57 – 31:58

piece of paper.

31:58 – 31:59

Even a spreadsheet, it's a

31:59 – 32:00

spreadsheet that one person

32:00 – 32:01

looks at instead of 10.

32:01 – 32:02

Yeah.

32:01 – 32:03

Yeah. And so it just totally

32:03 – 32:05

changed the dynamic, uh, of what

32:05 – 32:07

people's ownership was of the,

32:07 – 32:07

of the farm.

32:08 – 32:09

Um, and

32:10 – 32:11

it became a teaching tool.

32:11 – 32:12

And I think that's

32:12 – 32:14

part of part of our joy of our

32:14 – 32:15

farm is like the teaching

32:15 – 32:16

element.

32:18 – 32:19

So you've been working on this

32:19 – 32:21

now for the last three years.

32:21 – 32:23

How has this shifted and how

32:23 – 32:23

has,

32:23 – 32:25

how have things changed on your

32:25 – 32:26

operation as a result of it?

32:27 – 32:29

Yeah. So we

32:29 – 32:30

got to the end of the first

32:30 – 32:31

season and we realized that

32:31 – 32:32

everything up on that

32:32 – 32:33

whiteboard,

32:33 – 32:34

we were going to have to type

32:34 – 32:35

up.

32:34 – 32:36

for our organic certification.

32:37 – 32:39

And so previously what we were

32:39 – 32:40

doing was just having a giant

32:40 – 32:43

paper, an office supply paper

32:43 – 32:44

calendar where we're scribbling,

32:45 – 32:46

today we did this.

32:46 – 32:47

And then that was part of our

32:47 – 32:48

certification process.

32:49 – 32:50

Totally fine, totally legit,

32:51 – 32:52

works for record keeping, is

32:52 – 32:53

accepted by the USDA for that

32:53 – 32:54

process,

32:54 – 32:55

but kind of labor intensive.

32:56 – 32:57

And if

32:57 – 32:58

we wanted to go back and refer

32:58 – 32:59

like, hey, what did we do in

32:59 – 33:00

2012

33:00 – 33:02

when we had similar

33:02 – 33:03

circumstances?

33:03 – 33:04

How could we reference that?

33:07 – 33:08

It became clear like this

33:08 – 33:09

physical whiteboard

33:09 – 33:10

solved the initial problem, but

33:10 – 33:12

there was a natural evolution to

33:12 – 33:13

maybe a digital tool.

33:14 – 33:15

Could, could we build a piece of

33:15 – 33:16

software that functioned like

33:16 – 33:18

this whiteboard that could

33:18 – 33:19

assign tasks to team members?

33:19 – 33:21

We could pull in weather so we

33:21 – 33:23

could geotag it and have weather

33:23 – 33:24

pulled up automatically.

33:25 – 33:26

Could we have multi -year

33:26 – 33:27

references so we could look back

33:27 – 33:28

two or three years in similar

33:28 – 33:30

circumstances and see what

33:30 – 33:31

worked or what didn't?

33:32 – 33:33

Could we have details about who

33:33 – 33:35

did the work? So if it

33:35 – 33:36

looked

33:36 – 33:37

like the.

33:37 – 33:38

cultivator was set really well,

33:39 – 33:39

we could all learn from that.

33:42 – 33:44

And so I started the process of

33:44 – 33:45

developing FarmFluent, a piece

33:45 – 33:46

of software.

33:46 – 33:47

And that's been the last three

33:47 – 33:48

years journey.

33:48 – 33:49

I am a non -technical person.

33:49 – 33:52

I have no computer bones in my

33:52 – 33:52

body.

33:53 – 33:53

And so it's been a learning

33:53 – 33:54

process there.

33:55 – 33:57

The first year I hired a team on

33:57 – 33:59

Upworks, which is a

33:59 – 34:01

online gig economy tool.

34:02 – 34:04

I hired a team in India to build

34:04 – 34:04

the software. I thought, yeah,

34:04 – 34:07

10 grand, I can build this

34:07 – 34:07

thing. That's going to be

34:07 – 34:08

amazing.

34:08 – 34:09

It'll change the world.

34:10 – 34:11

How hard can it be?

34:14 – 34:15

Boy, was I wrong.

34:17 – 34:18

So between the barriers of like

34:18 – 34:19

language,

34:20 – 34:21

farming context, you know, the

34:21 – 34:22

farming context in India versus

34:22 – 34:24

a Midwest grain farmer, wildly

34:24 – 34:24

different.

34:25 – 34:26

And really just being non

34:26 – 34:27

-technical, just there's a whole

34:27 – 34:29

language that's spoken about how

34:29 – 34:31

do you develop software and what

34:31 – 34:32

I might visualize in my mind

34:32 – 34:33

communicating that to a team

34:33 – 34:34

was,

34:34 – 34:35

it was just amazing.

34:35 – 34:36

So I spent a year,

34:37 – 34:38

a chunk of money,

34:39 – 34:40

and I built a bicycle that

34:40 – 34:42

didn't have wheels or a seat on

34:42 – 34:43

it.

34:43 – 34:44

It kind of looked like the

34:44 – 34:45

thing, but it was not going to

34:45 – 34:46

work.

34:47 – 34:49

Yeah. And so we restarted the

34:49 – 34:53

process here in 2024 with

34:53 – 34:54

some grant funding.

34:54 – 34:56

So the project today has a grant

34:56 – 34:58

-funded software tool, and I'm

34:58 – 34:58

working with a team in

34:58 – 34:59

Minneapolis.

34:59 – 35:01

So a US -based team.

35:01 – 35:03

I have a product manager who

35:03 – 35:06

helped build a large

35:06 – 35:08

software tool that vegetable

35:08 – 35:08

farmers use.

35:09 – 35:10

And we're working with four or

35:10 – 35:12

five farmers here in the US to

35:12 – 35:13

kind of test it out, to learn

35:13 – 35:15

from what they need in their

35:15 – 35:17

farms. And so we're restarting,

35:17 – 35:18

we're relearning,

35:18 – 35:20

but we're building a tool that

35:20 – 35:22

takes that whiteboard concept

35:22 – 35:24

and starts applying it

35:24 – 35:25

to a digital world.

35:25 – 35:26

And eventually we're thinking

35:26 – 35:27

about AI.

35:27 – 35:28

So I'm excited to talk with you

35:28 – 35:29

about AI too.

35:30 – 35:31

Yeah, I want to dig into the AI

35:31 – 35:33

conversation, but I

35:35 – 35:36

want to not lose sight of my

35:36 – 35:37

original question quite yet,

35:37 – 35:39

which is what's changed on your

35:39 – 35:40

farm in the last three years

35:41 – 35:44

as what what in

35:45 – 35:47

terms of crop responses or crop

35:47 – 35:48

management, you spoke about this

35:48 – 35:50

being kind of the nexus point

35:50 – 35:51

for a conversation and for

35:51 – 35:53

facilitating discussion around

35:53 – 35:54

different points of view, which

35:54 – 35:55

which is awesome.

35:55 – 35:58

But has that has that

35:58 – 36:01

resulted in actually improving

36:01 – 36:02

yield, which was your original

36:02 – 36:03

question?

36:03 – 36:05

Yeah, it has improved yield.

36:08 – 36:09

The first year that we used the

36:09 – 36:11

FarmFill board, we had a

36:12 – 36:15

record yield up to that point.

36:16 – 36:17

Of course, weather,

36:17 – 36:18

fertility, and everything else

36:19 – 36:20

has a large role to play.

36:20 – 36:23

I would be very, very remiss to

36:23 – 36:24

say that if you put a 4x8

36:24 – 36:26

whiteboard in your farm shop,

36:26 – 36:26

you're going to have the best

36:26 – 36:27

yields ever.

36:29 – 36:30

That simply is not true.

36:30 – 36:32

true, folks. That will not

36:32 – 36:32

happen.

36:35 – 36:37

What has changed is that

36:37 – 36:40

the ownership of the operation

36:40 – 36:41

has changed. I think as a result

36:41 – 36:42

of that,

36:42 – 36:43

we have better execution day -to

36:43 – 36:44

-day. As a result of better

36:44 – 36:46

execution day -to -day, we have

36:46 – 36:47

more consistent yields.

36:49 – 36:52

What it is doing is reducing our

36:53 – 36:54

lived experience of stress.

36:56 – 36:59

What it is doing is taking the

36:59 – 37:00

record keeping process and

37:00 – 37:01

making that much simpler year to

37:01 – 37:02

year.

37:03 – 37:03

And

37:03 – 37:04

this

37:05 – 37:06

will be our third

37:06 – 37:08

season with the tool in place.

37:10 – 37:12

So I'd have to look at 25 years

37:12 – 37:14

and say, can I chart that out?

37:14 – 37:15

Hard to say exactly.

37:16 – 37:17

Again, weather,

37:17 – 37:19

fertility, crops, everything

37:19 – 37:20

else comes into play

37:20 – 37:21

significantly.

37:21 – 37:24

But it's changed the day -to

37:24 – 37:25

-day experience on our farm

37:25 – 37:27

significantly and helped our

37:27 – 37:28

team.

37:31 – 37:31

All right,

37:31 – 37:32

let's talk about AI.

37:33 – 37:34

You've mentioned FieldArk twice

37:34 – 37:35

now,

37:35 – 37:36

and I have

37:37 – 37:38

conversations with people all

37:38 – 37:41

the time who are using AI for a

37:41 – 37:43

variety of purposes and in

37:43 – 37:45

usually partially delighted and

37:45 – 37:47

partially terrified and

37:47 – 37:48

partially,

37:50 – 37:51

yeah, there's just a mix of

37:51 – 37:53

emotions about the results and

37:53 – 37:54

the performance that they're

37:54 – 37:55

getting from it.

37:55 – 37:57

There's delight and

37:57 – 37:59

occasional frustration in equal

37:59 – 38:00

measure, particularly from some

38:00 – 38:01

of the more mainstream tools.

38:02 – 38:04

What have you been playing

38:04 – 38:04

around with?

38:05 – 38:06

I did have an opportunity to use

38:06 – 38:07

Fieldark when it launched.

38:07 – 38:09

I was excited to see it.

38:09 – 38:10

We talked earlier about

38:10 – 38:12

agronomists. You were one of the

38:12 – 38:13

original beta users.

38:14 – 38:15

Maybe. Maybe I signed up, yeah.

38:17 – 38:19

We have a great,

38:20 – 38:21

long -time, conventional

38:21 – 38:23

agriculture agronomist who I

38:23 – 38:25

call the Hope Kitsuo

38:25 – 38:25

samples.

38:26 – 38:28

And if something seems really

38:28 – 38:29

confusing, I might call them or

38:29 – 38:30

send them a photo.

38:30 – 38:31

But beyond that,

38:32 – 38:32

we are on our own.

38:33 – 38:36

And so I uploaded all of our

38:36 – 38:38

soil test results into Fieldark,

38:38 – 38:39

and I started asking

38:39 – 38:40

questions like,

38:40 – 38:41

okay, let's, let's walk through

38:41 – 38:42

this. What can you do?

38:42 – 38:45

And I had a more robust

38:45 – 38:46

conversation with field lark

38:46 – 38:48

than I've had with my local

38:48 – 38:49

agronomist over the last six,

38:49 – 38:51

seven years, uh, in about 30

38:51 – 38:52

minutes.

38:52 – 38:53

So in this way, it was really

38:53 – 38:54

exciting.

38:54 – 38:55

Um,

38:55 – 38:57

Out of curiosity, what was the

38:57 – 38:58

nature of the questions that you

38:58 – 38:59

were asking and what was,

38:59 – 39:00

um,

39:00 – 39:01

what were the types of responses

39:01 – 39:02

that you were getting back?

39:04 – 39:05

Um, I started out, I've just

39:05 – 39:06

pointed up here.

39:06 – 39:07

I started out very generic.

39:07 – 39:08

So I said, attach it

39:09 – 39:10

to my soil samples.

39:10 – 39:11

What can you tell me about my

39:11 – 39:12

farm?

39:12 – 39:13

I got a long response, and then

39:13 – 39:14

we started digging into it.

39:14 – 39:15

So then I started saying,

39:16 – 39:16

39:18 – 39:20

our farm, we have some peat

39:20 – 39:21

soils, which are really high in

39:21 – 39:22

organic matter.

39:22 – 39:24

And so that was throwing off

39:24 – 39:25

some of the model.

39:24 – 39:25

And so I said, OK, remove the

39:25 – 39:27

fields with super, super high

39:27 – 39:28

organic matter.

39:28 – 39:30

Let's dig into some

39:30 – 39:33

of my target yield goals.

39:35 – 39:36

What can you tell about

39:36 – 39:37

micronutrients?

39:37 – 39:39

What suggestions would you have

39:39 – 39:40

for increasing these results?

39:43 – 39:44

I started asking about foliar

39:44 – 39:46

boron and zinc on corn.

39:47 – 39:48

And, you

39:49 – 39:50

know, got interesting answers

39:50 – 39:51

there.

39:51 – 39:51

39:51 – 39:53

it's

39:54 – 39:56

been fun. I think my experience

39:56 – 39:57

with AI has been,

39:58 – 39:59

I think about it maybe as like a

39:59 – 40:02

college intern, where it

40:02 – 40:04

has access to knowledge,

40:05 – 40:07

but not quite

40:08 – 40:10

the context to offer

40:12 – 40:15

original insights that it can

40:15 – 40:15

assemble

40:15 – 40:17

and answer

40:18 – 40:19

things

40:20 – 40:21

you maybe give a framework for,

40:21 – 40:23

but it's not generating, for me

40:23 – 40:24

at least,

40:24 – 40:25

totally

40:27 – 40:28

new and original concepts.

40:29 – 40:31

The temptation in my experience,

40:31 – 40:32

and this is, I'm talking about

40:32 – 40:34

chat GPT or other large language

40:34 – 40:35

models, not necessarily field

40:35 – 40:36

work, but

40:37 – 40:39

my experience with AI is this,

40:39 – 40:41

maybe this same sort of delight

40:41 – 40:43

and distrust that

40:43 – 40:45

initially it's really tempting

40:45 – 40:46

to interact with it and just

40:46 – 40:47

think that it's going to do the

40:47 – 40:48

thinking for you.

40:48 – 40:50

But where I found AI to be most

40:50 – 40:52

helpful is if you come to the

40:52 – 40:53

tool with a question,

40:54 – 40:55

with context,

40:56 – 40:57

and have done the work of

40:57 – 40:58

thinking before,

40:59 – 41:00

it's helpful.

41:00 – 41:02

If you're approaching the thing

41:02 – 41:04

with not much thought or without

41:04 – 41:07

much inquiry or kind of

41:08 – 41:09

context,

41:09 – 41:10

the

41:10 – 41:11

results you get back are

41:11 – 41:12

mediocre.

41:14 – 41:17

I think there's a fundamental

41:17 – 41:17

difference, obviously.

41:19 – 41:21

a great deal about the

41:21 – 41:22

development of AI and the work

41:22 – 41:24

that we've put into FieldArk.

41:26 – 41:27

And there's a fundamental

41:27 – 41:29

difference between the way

41:29 – 41:32

humans respond to answering

41:32 – 41:33

questions and

41:34 – 41:36

the way a machine responds.

41:38 – 41:39

And there are

41:39 – 41:41

two key ones that I'll focus on

41:41 – 41:42

for the moment.

41:42 – 41:43

One of them is,

41:45 – 41:46

OK, you're an expert organic

41:46 – 41:48

corn and soybean farmer.

41:49 – 41:50

And for your

41:50 – 41:51

local environment, you've got, I

41:51 – 41:53

mean, there's no question you

41:53 – 41:54

know way more about that than I

41:54 – 41:55

do.

41:55 – 41:57

And I ask you the question to

41:57 – 41:59

say, hey, I want to grow organic

41:59 – 42:00

corn.

42:03 – 42:05

How many cultivation

42:07 – 42:08

passes do I need to make for

42:08 – 42:09

effective weed control?

42:10 – 42:12

And you're going to say, well,

42:12 – 42:13

it depends on this and that.

42:13 – 42:14

And the other thing, it depends

42:14 – 42:15

on what you're doing.

42:15 – 42:16

It depends on there's there's a

42:16 – 42:18

whole bunch of variables like

42:18 – 42:19

you. You can't give an answer to

42:19 – 42:22

that question without knowing a

42:22 – 42:23

whole bunch of context.

42:23 – 42:24

So knowing the context is

42:24 – 42:26

important that any human is

42:26 – 42:27

going to say, well, I don't know

42:27 – 42:28

the answer to that.

42:28 – 42:29

It's like there is no cut and

42:29 – 42:30

dried answer.

42:30 – 42:32

Yeah. And we will respond.

42:32 – 42:34

Humans will respond by asking

42:34 – 42:36

for more context or describing

42:36 – 42:37

that more context is needed.

42:38 – 42:38

That's one

42:39 – 42:40

significant

42:41 – 42:43

And the other significant

42:43 – 42:45

difference is we'll ask about

42:45 – 42:45

one thing,

42:46 – 42:49

and for someone who is an expert

42:49 – 42:51

or has experience and knowledge

42:51 – 42:52

in a certain area,

42:53 – 42:54

they will say, well, actually,

42:55 – 42:56

while that might appear to be a

42:56 – 42:57

good question,

42:57 – 42:58

you're actually completely

42:58 – 42:59

missing this thing over here.

43:00 – 43:00

And what you should be thinking

43:00 – 43:01

about is you shouldn't be

43:01 – 43:03

thinking about X, you should be

43:03 – 43:04

thinking about Y.

43:05 – 43:06

The machine can't do that,

43:06 – 43:09

doesn't do that, or does it very

43:09 – 43:09

poorly at this point.

43:11 – 43:13

And so I

43:15 – 43:16

had this experience

43:19 – 43:19

I've

43:20 – 43:22

had some of these interviews,

43:22 – 43:23

both here on the podcast and in

43:23 – 43:24

person,

43:25 – 43:26

that

43:26 – 43:28

haven't been publicly released

43:28 – 43:28

yet,

43:29 – 43:30

where I interviewed people from

43:30 – 43:32

a military background about

43:32 – 43:34

research work that they did

43:34 – 43:36

decades ago on plant pathology

43:36 – 43:37

and so forth.

43:38 – 43:40

And when you get into some of

43:40 – 43:42

those conversations that have

43:42 – 43:44

national security implications,

43:46 – 43:48

the answers become very direct

43:48 – 43:49

and concise.

43:49 – 43:51

They become yes, no.

43:51 – 43:53

It's like they will only answer

43:53 – 43:55

the question that you asked and

43:55 – 43:56

they don't elaborate.

43:57 – 43:59

And so in order to have any type

43:59 – 44:00

of conversation or any type of

44:00 – 44:02

dialogue, you have to be smart

44:02 – 44:03

enough to

44:03 – 44:04

know what questions to ask.

44:05 – 44:06

Because if you don't know the

44:06 – 44:07

questions, they're not going to

44:07 – 44:08

volunteer the information.

44:09 – 44:10

And

44:11 – 44:13

think the same thing is so true

44:13 – 44:15

of these AI tools like

44:15 – 44:16

FieldLark,

44:17 – 44:20

is you have to, to some

44:20 – 44:21

extent, you have to be smart

44:21 – 44:22

enough to know the question to

44:22 – 44:26

ask. And this is one of the

44:26 – 44:27

things that I've tried to

44:27 – 44:28

develop in thinking about,

44:29 – 44:31

how do we develop these tools to

44:31 – 44:33

help people become better, to

44:33 – 44:35

help people ask better questions

44:35 – 44:35

and so forth?

44:36 – 44:38

is in field work we have this

44:38 – 44:40

these follow -up questions um

44:40 – 44:42

and as part of each response and

44:42 – 44:44

it is that is our

44:44 – 44:46

initial attempt

44:46 – 44:48

to to

44:48 – 44:50

inspire more of a dialogue and

44:50 – 44:52

to help all of us ask better

44:52 – 44:53

questions and remind us of

44:53 – 44:54

questions that we might perhaps

44:54 – 44:55

should be thinking about but

44:55 – 44:56

that we're not thinking about

44:56 – 44:57

right now

44:57 – 44:58

yep john

44:59 – 45:00

as you've been building this

45:00 – 45:01

tool

45:01 – 45:03

how do you how do you balance

45:03 – 45:04

the

45:04 – 45:06

what I might perceive as risks

45:06 – 45:07

of AI, you know, either energy

45:07 – 45:10

consumption or data privacy or

45:10 – 45:10

even

45:11 – 45:12

kind of intellectual, what

45:12 – 45:13

you're describing, kind of

45:13 – 45:15

intellectual risk of keeping

45:16 – 45:17

in the creativity of the human

45:17 – 45:19

element. How do you balance that

45:19 – 45:20

with the

45:20 – 45:21

opportunities or the

45:23 – 45:25

utility of the tools that you're

45:25 – 45:26

building?

45:27 – 45:28

Well,

45:27 – 45:28

there's

45:31 – 45:32

several questions within the

45:32 – 45:34

question that you've asked, but

45:34 – 45:36

I'd say broadly I've taken a

45:36 – 45:38

fairly pragmatic approach to

45:38 – 45:38

just say

45:42 – 45:44

AI is what we're currently

45:44 – 45:46

calling AI, which is largely

45:46 – 45:47

defined in terms of LLMs, these

45:47 – 45:48

large language models.

45:48 – 45:49

And this is obviously this whole

45:49 – 45:51

space is rapidly iterating and

45:51 – 45:52

evolving.

45:52 – 45:55

But there's a very pragmatic

45:55 – 45:57

element of saying, this is going

45:57 – 45:59

to exist whether we create

45:59 – 46:00

FieldArk or not.

46:01 – 46:03

This whole space is like,

46:05 – 46:06

the guardrails are off

46:06 – 46:08

and we don't have control of

46:08 – 46:09

this horse anymore.

46:10 – 46:13

So there is an element of

46:15 – 46:17

if you can't beat them, you

46:17 – 46:19

might as well join them and

46:19 – 46:21

let's create something that is

46:21 – 46:23

better than what currently

46:23 – 46:24

exists for

46:24 – 46:26

a specific use case.

46:26 – 46:27

Let's create something that is

46:27 – 46:29

better, that helps people

46:29 – 46:30

better, that helps farmers and

46:30 – 46:31

agronomists to be more effective

46:31 – 46:32

and

46:32 – 46:33

that has significantly more

46:33 – 46:36

powerful for a specific use case

46:36 – 46:39

than ChatGPT or Grok or Gemini

46:39 – 46:40

or any of these other tools that

46:40 – 46:41

are out there.

46:42 – 46:45

So that's the one consideration.

46:45 – 46:47

And then another aspect of the

46:47 – 46:48

question you asked about energy

46:48 – 46:50

consumption, this kind of

46:50 – 46:53

relates to my initial response,

46:53 – 46:54

46:54 – 46:55

46:59 – 47:00

are perfectly comfortable in

47:01 – 47:04

many domains of life using

47:04 – 47:06

significant energy

47:07 – 47:10

And you could argue using energy

47:10 – 47:12

wastefully for outcomes that we

47:12 – 47:14

perceive to be valuable.

47:15 – 47:16

So, for example,

47:16 – 47:17

I'm trying to forget or trying

47:17 – 47:19

to recall who did this analysis.

47:19 – 47:21

I want to say it was Wendell

47:21 – 47:22

Berry, but I don't think it was

47:22 – 47:23

Wendell Berry.

47:25 – 47:26

Wendell might have talked about

47:26 – 47:27

it, though, in Unsettling

47:27 – 47:28

America,

47:28 – 47:30

but something to the effect that

47:30 – 47:32

our modern agriculture today,

47:33 – 47:36

we use three to four calories of

47:37 – 47:38

petroleum energy to produce a

47:38 – 47:40

single calorie of food energy,

47:40 – 47:42

which is not sustainable on its

47:42 – 47:43

face.

47:44 – 47:46

It's working for a temporary

47:46 – 47:47

moment in history, but when you

47:47 – 47:50

look at overall human history in

47:50 – 47:52

terms of centuries or millennia,

47:52 – 47:54

then I think at some point we

47:54 – 47:55

will look backward and say that

47:55 – 47:57

was a very wasteful use of

47:57 – 47:58

energy.

48:00 – 48:01

And similarly,

48:02 – 48:04

so we have to look at these

48:05 – 48:08

LLMs have been created,

48:08 – 48:09

what are we going to use them

48:09 – 48:10

for?

48:11 – 48:12

How are we using that energy?

48:13 – 48:14

And so I would argue that using

48:14 – 48:15

that energy for

48:16 – 48:18

the purposes of facilitating a

48:18 – 48:19

48:19 – 48:21

a regenerative agriculture is a

48:21 – 48:22

valuable use of that energy,

48:22 – 48:24

much more valuable than what

48:24 – 48:25

it's probably most widely being

48:25 – 48:27

used for on a consumer audience

48:27 – 48:27

perspective.

48:28 – 48:30

Yeah. Yeah. Looking up a recipe

48:30 – 48:31

for spaghetti.

48:31 – 48:32

Exactly.

48:32 – 48:33

Probably not.

48:34 – 48:34

Not

48:35 – 48:35

helping the world.

48:36 – 48:37

Yeah.

48:36 – 48:37

Yeah.

48:37 – 48:38

Well, it's

48:40 – 48:43

one of the principles that I

48:43 – 48:44

find is true at our farm is that

48:44 – 48:46

we've we we have done well by

48:46 – 48:47

being early adopters,

48:48 – 48:49

by being willing to ask

48:49 – 48:50

questions, by being able to lean

48:50 – 48:51

into things, whether that was

48:52 – 48:54

organic or certain technologies,

48:54 – 48:56

say a GPS for

48:56 – 48:57

tractors being able to drive

48:57 – 48:58

straight so you could cultivate

48:58 – 49:00

better, whatever, whatever it

49:00 – 49:01

may be. We've, we've generally

49:01 – 49:03

leaned into being early

49:03 – 49:04

adopters. And so that's been my

49:04 – 49:05

approach to this

49:06 – 49:07

tool in question too, is that

49:07 – 49:08

there

49:10 – 49:11

needs to be a healthy amount of

49:11 – 49:12

skepticism. There are always

49:12 – 49:13

trade -offs.

49:15 – 49:17

And at the same time, it is, it

49:17 – 49:18

is a tool.

49:21 – 49:22

And

49:23 – 49:24

we,

49:27 – 49:28

we know that

49:29 – 49:31

commodity agriculture is going

49:31 – 49:32

to be using these things too.

49:33 – 49:33

And.

49:34 – 49:35

and it will be to

49:36 – 49:37

kind of perpetuate those

49:37 – 49:38

systems.

49:39 – 49:40

And if we're not thinking about

49:40 – 49:42

how we're going to participate

49:42 – 49:43

or respond, I think

49:45 – 49:45

those

49:46 – 49:47

conventional systems will

49:48 – 49:49

continue to dominate.

49:51 – 49:53

So yeah, I would say I kind of

49:53 – 49:54

take a similar perspective about

49:55 – 49:56

using the tools for,

49:58 – 50:01

you know, for good, for driving

50:01 – 50:02

change.

50:02 – 50:03

Well, in the way that we've

50:03 – 50:05

developed FieldLark, I'm

50:05 – 50:06

actually not thinking about it

50:06 – 50:09

in terms of conventional

50:09 – 50:11

systems versus organic or

50:11 – 50:12

regenerative systems.

50:13 – 50:14

We built it,

50:15 – 50:16

I built it very deliberately

50:17 – 50:19

to be

50:21 – 50:23

what's the word that I'm looking

50:23 – 50:25

for, to be practiced agnostic or

50:25 – 50:26

to be,

50:26 – 50:29

but to instead function from a

50:29 – 50:31

first principles perspective and

50:31 – 50:32

to function from an outcomes

50:32 – 50:33

perspective.

50:34 – 50:35

So, and when I say an outcomes

50:35 – 50:36

perspective,

50:38 – 50:39

FieldLark is directly,

50:40 – 50:42

is directed to take second and

50:42 – 50:43

third order consequences into

50:43 – 50:45

consideration as well to the

50:45 – 50:46

degree that they're understood

50:46 – 50:47

or known.

50:48 – 50:50

And so it's quite fascinating.

50:51 – 50:53

I spent quite some time thinking

50:53 – 50:54

about

50:54 – 50:57

in developing the

50:57 – 50:59

ethical framework

50:59 – 51:02

of what's the ethical framework

51:02 – 51:03

and what are the first

51:03 – 51:04

principles framework that

51:04 – 51:06

FieldLARC is to operate within

51:06 – 51:06

and

51:06 – 51:08

what are its foundational

51:08 – 51:09

directives.

51:09 – 51:09

And so when you,

51:11 – 51:12

and it's interesting because we

51:12 – 51:13

didn't give FieldLARC any

51:13 – 51:14

directives

51:15 – 51:18

to, for example, to say don't

51:18 – 51:19

make recommendations for

51:19 – 51:21

pesticides or don't make

51:21 – 51:22

recommendations for anhydrous

51:22 – 51:24

ammonia or a variety of

51:24 – 51:25

different things that I could

51:25 – 51:25

point out.

51:26 – 51:28

But what is intriguing is that

51:28 – 51:30

FieldLark will not make,

51:31 – 51:33

at least not readily on an

51:33 – 51:34

initial request, it will not

51:34 – 51:36

make recommendations for those

51:36 – 51:39

types of things because of its

51:39 – 51:41

ethical framework and its first

51:41 – 51:42

principles framework thinking

51:42 – 51:44

where it takes into

51:44 – 51:45

consideration second

51:46 – 51:47

and third order consequences.

51:48 – 51:49

Does all of a sudden say, well,

51:49 – 51:51

actually, yes, anhydrous ammonia

51:51 – 51:52

can be a very valuable, useful

51:52 – 51:55

form of nitrogen from a crop and

51:55 – 51:56

agronomy perspective,

51:56 – 51:58

but anhydrous ammonia also has

51:58 – 52:00

all of these other

52:00 – 52:01

externalities

52:02 – 52:03

that are perceived to be

52:03 – 52:05

unhealthy to the environment and

52:05 – 52:06

to the ecosystem.

52:06 – 52:06

And therefore,

52:07 – 52:08

FieldLark, without being

52:08 – 52:09

specifically told to,

52:10 – 52:12

it will selectively, it will

52:12 – 52:14

make preferential

52:14 – 52:15

recommendations for other forms

52:15 – 52:16

of nitrogen

52:17 – 52:18

other than anhydrous ammonia.

52:18 – 52:20

So it's not that it doesn't make

52:20 – 52:21

recommendations for nitrogen,

52:21 – 52:24

it's just it's it has it has

52:24 – 52:26

that first principles approach,

52:26 – 52:28

which is so intriguing because

52:28 – 52:30

we deliberately asked it,

52:31 – 52:33

directed Fieldmark to have no

52:33 – 52:35

biases, to remove all biases in

52:35 – 52:38

favor of any system.

52:38 – 52:39

It's not in favor of organic, or

52:39 – 52:41

not in favor of regenerative, or

52:41 – 52:41

not in favor of

52:42 – 52:44

of contemporary agricultural

52:44 – 52:46

systems, but instead to operate

52:46 – 52:47

exclusively from a first

52:47 – 52:48

principles perspective and not

52:48 – 52:49

take any biases into

52:49 – 52:50

consideration.

52:50 – 52:52

And that very fact makes it

52:52 – 52:53

appear to some people as if

52:53 – 52:54

though it were biased.

52:55 – 52:56

Yeah.

52:58 – 52:59

Challenging our own worldviews.

53:00 – 53:00

Yeah. Yeah.

53:01 – 53:01

Yeah.

53:02 – 53:05

When you reflect on

53:05 – 53:07

what you've learned in the last

53:07 – 53:09

couple of months of releasing

53:09 – 53:10

FieldArk, what's been

53:10 – 53:11

unexpected?

53:16 – 53:18

That's a good question.

53:19 – 53:20

53:23 – 53:24

would say

53:25 – 53:27

perhaps this

53:29 – 53:30

is a reflection of my own

53:30 – 53:31

biases.

53:31 – 53:32

53:33 – 53:35

you described, I tend to be a

53:35 – 53:37

technology early adopter.

53:37 – 53:39

I've leaned in very strongly

53:39 – 53:40

into

53:41 – 53:43

testing these various GPT

53:43 – 53:45

platforms when they were first

53:45 – 53:45

launched.

53:46 – 53:48

I've been thinking about them

53:48 – 53:50

for over a decade.

53:51 – 53:53

I had some early exposure to the

53:53 – 53:54

technology that became these

53:54 – 53:55

large language models long

53:55 – 53:56

before they were publicly

53:56 – 53:57

released.

53:57 – 53:58

And so I've been thinking about

53:58 – 53:59

this for some time, and I've

53:59 – 54:01

leaned heavily into using them

54:01 – 54:05

and building them to work for

54:05 – 54:05

me.

54:06 – 54:07

And the

54:08 – 54:08

piece that

54:09 – 54:11

I'm perhaps the most surprised

54:11 – 54:12

about

54:12 – 54:14

is that many people,

54:16 – 54:18

many users of FieldLark, and

54:18 – 54:20

based on what I can see, I'm

54:20 – 54:22

assuming other GPT models as

54:22 – 54:23

well,

54:24 – 54:25

are still using it as a

54:25 – 54:26

glorified Google search.

54:27 – 54:28

They're asking more specific

54:28 – 54:29

questions,

54:30 – 54:33

but they haven't

54:33 – 54:35

really understood the

54:35 – 54:37

implications of what this

54:37 – 54:38

technology can do for them once

54:38 – 54:39

it is trained for them.

54:40 – 54:41

And you mentioned this at the

54:41 – 54:42

beginning of the con, you

54:42 – 54:44

uploaded a bunch of

54:46 – 54:47

soil samples. But what you can

54:47 – 54:49

actually do is within FieldLark,

54:49 – 54:50

you have these different

54:50 – 54:51

projects.

54:51 – 54:53

So you can have a project be a

54:53 – 54:55

specific field or a farm or a

54:55 – 54:55

specific crop that you're

54:55 – 54:58

working on. Or let's just say

54:58 – 54:59

you use it for your whole farm.

54:59 – 55:00

55:00 – 55:03

you can write up a document or

55:03 – 55:05

even even better.

55:05 – 55:06

And this is the way that I this

55:06 – 55:09

is the way that I use FieldLark

55:09 – 55:10

and other tools as well.

55:10 – 55:12

When I'm building out an initial

55:12 – 55:13

profile, I won't build a

55:13 – 55:15

profile. I'll use the AI tool to

55:15 – 55:16

help me build a profile.

55:17 – 55:18

So I'll say,

55:19 – 55:20

hey, Fieldmark,

55:20 – 55:23

I am farming at this location

55:23 – 55:25

with these crops,

55:26 – 55:28

with these soil types, here are

55:28 – 55:29

my soil characteristics, here

55:29 – 55:30

are my dominant weed species.

55:31 – 55:34

Help me build a profile that

55:36 – 55:38

describes my context that I can

55:38 – 55:39

use to train an AI tool.

55:41 – 55:42

And you'll go back and forth,

55:42 – 55:43

you iterate half a dozen times,

55:44 – 55:45

and I end up with

55:45 – 55:48

usually as much as a 20 to a 30

55:48 – 55:49

-page document that describes my

55:49 – 55:51

context in detail.

55:51 – 55:53

These are the common weather

55:53 – 55:54

conditions,

55:54 – 55:56

here's the climate, here's the

55:56 – 55:57

USDA growing zones,

55:57 – 55:58

here's all this.

55:58 – 56:00

And let the AI actually help

56:00 – 56:01

build out that context rather

56:01 – 56:02

than you trying to do it

56:02 – 56:03

yourself. And then you take that

56:03 – 56:05

information and you upload it

56:05 – 56:06

into

56:06 – 56:08

a project, put it into a

56:08 – 56:09

document, put it into that

56:09 – 56:09

project,

56:10 – 56:11

and now it

56:11 – 56:13

will constantly reference back

56:13 – 56:14

to that training data set.

56:15 – 56:16

And you can constantly add to

56:16 – 56:17

that and build it.

56:17 – 56:19

And so what happens over time,

56:20 – 56:22

the information that you get

56:22 – 56:24

back becomes ever more

56:24 – 56:27

customized and focused on your

56:27 – 56:29

specific operation as you keep

56:29 – 56:30

building and as you keep

56:30 – 56:31

training it. And so the quality

56:31 – 56:32

of the answers and the quality

56:32 – 56:33

of the information you get back

56:33 – 56:34

out of it

56:35 – 56:36

becomes very valuable.

56:36 – 56:38

And the reality is what I'm

56:38 – 56:39

talking about in terms of

56:39 – 56:41

training these AIs to work for

56:41 – 56:42

you

56:42 – 56:45

is this can be as

56:45 – 56:46

simple as

56:46 – 56:48

30 minute exchange back and

56:48 – 56:50

forth. They are very good at

56:50 – 56:52

actually identifying what

56:52 – 56:54

information would be valuable

56:54 – 56:55

for them to understand your

56:55 – 56:56

context.

56:57 – 56:58

Yeah.

56:58 – 57:00

Where I've used that similarly

57:00 – 57:01

is with this FarmFlow project.

57:02 – 57:02

I mentioned it's been grant

57:02 – 57:03

funded.

57:03 – 57:04

I'm also not a grant writer.

57:05 – 57:07

So I've been using it to

57:07 – 57:09

apply for grants

57:10 – 57:12

by giving it context.

57:12 – 57:14

So like sharing previous grant

57:14 – 57:15

proposals or previous language,

57:16 – 57:17

57:17 – 57:19

documents, the request for

57:19 – 57:20

proposals,

57:20 – 57:21

you know, pitch decks, kind of

57:21 – 57:22

all,

57:22 – 57:24

everything. And I've created a

57:24 – 57:25

project in Chattopadhyay.

57:25 – 57:26

And then that,

57:26 – 57:28

you know, when I get off the

57:28 – 57:29

farm and tuck

57:29 – 57:31

kiddos into bed, and then it's

57:31 – 57:32

time to start working on a grant

57:32 – 57:34

at night, and I've got two brain

57:34 – 57:35

cells left.

57:36 – 57:38

It's been really helpful to kind

57:38 – 57:38

57:39 – 57:40

be that partner,

57:40 – 57:41

to kind of take it to that next

57:41 – 57:42

level.

57:43 – 57:44

One of the projects that I did,

57:44 – 57:45

just to give you an example,

57:47 – 57:49

this is actually one of the

57:49 – 57:50

projects that I did that I then

57:50 – 57:52

refused to ever use again,

57:52 – 57:55

just from a principle to

57:55 – 57:56

perspective.

57:57 – 57:58

I trained

57:59 – 58:01

an instance

58:01 – 58:02

of Gemini

58:05 – 58:07

to sound like Jon Kempf, to talk

58:07 – 58:09

like Jon Kempf, and to write

58:09 – 58:10

like Jon Kempf.

58:10 – 58:12

I said, OK, here are

58:12 – 58:14

50, here are the transcripts

58:14 – 58:16

from 50 webinars and dozens of

58:16 – 58:17

podcasts.

58:18 – 58:20

Here are a long list of articles

58:20 – 58:21

that I've written and blog posts

58:21 – 58:22

that I've written.

58:22 – 58:23

And so I trained it on all this

58:23 – 58:24

and I said,

58:24 – 58:26

analyze all

58:27 – 58:28

of this

58:28 – 58:30

and write a, I forget what the

58:30 – 58:33

proper term is, but kind

58:33 – 58:36

of a brand tone document of how

58:36 – 58:37

does Jon's

58:38 – 58:41

language and his expressions and

58:41 – 58:43

write a document to train an AI

58:43 – 58:45

so that it speaks and writes

58:45 – 58:46

like John Kempf.

58:47 – 58:49

And it generated like a 50 -page

58:49 – 58:51

document with this extensive

58:51 – 58:53

detail about the tones and the

58:53 – 58:54

type of voice inflections.

58:54 – 58:56

It analyzed all of it, wrote up

58:56 – 58:57

this document,

58:58 – 59:00

and I provided the document as a

59:00 – 59:02

training tool, as a training

59:02 – 59:02

set.

59:03 – 59:05

And it came back and I said, all

59:05 – 59:06

right,

59:06 – 59:08

behave as John Kempf and write

59:08 – 59:09

an article on quorum sensing.

59:10 – 59:11

And I looked at that article

59:11 – 59:13

like, this is freaky.

59:14 – 59:15

It's done.

59:17 – 59:18

Well,

59:18 – 59:19

it's

59:19 – 59:20

not that

59:21 – 59:22

59:22 – 59:25

concern was not that it sounded

59:25 – 59:26

too much like me.

59:27 – 59:29

What I was concerned about is

59:29 – 59:30

that if you start using it,

59:30 – 59:32

tools, it's just very similar to

59:32 – 59:34

once

59:34 – 59:37

maps apps were prevalent on

59:37 – 59:38

every phone, people lost their

59:38 – 59:39

sense of navigation.

59:40 – 59:41

They can no longer,

59:41 – 59:42

people generally, there are

59:42 – 59:43

exceptions of course, but people

59:43 – 59:44

generally no longer have a good

59:44 – 59:45

sense of direction or a good

59:45 – 59:46

ability to navigate.

59:47 – 59:50

And you start using tools like

59:50 – 59:52

that for writing articles and so

59:52 – 59:53

forth, you'll start losing your

59:53 – 59:55

ability to think critically.

59:55 – 59:57

And so there are

59:59 – 1:00:00

ways I think

1:00:01 – 1:00:03

we need to think carefully about

1:00:03 – 1:00:04

how do we use these tools like

1:00:04 – 1:00:06

in the way that we've built

1:00:06 – 1:00:07

FieldArc, I've thought a lot

1:00:07 – 1:00:09

about how do we use this to not

1:00:09 – 1:00:10

replace people, but to make

1:00:10 – 1:00:11

people better.

1:00:12 – 1:00:14

And there's only so much of that

1:00:14 – 1:00:16

we can do as a tool designer.

1:00:17 – 1:00:18

Another part of that has to be

1:00:18 – 1:00:20

picked up by the tool user and

1:00:20 – 1:00:21

the tool user has to say, all

1:00:21 – 1:00:22

right,

1:00:22 – 1:00:24

I want this to help me do the

1:00:24 – 1:00:26

things that I'm not great at.

1:00:26 – 1:00:27

It can help really expedite

1:00:27 – 1:00:29

research, for example, or to

1:00:29 – 1:00:30

learn things that I didn't

1:00:30 – 1:00:31

already know.

1:00:31 – 1:00:32

But then when I learned those

1:00:32 – 1:00:34

things, there is still this

1:00:34 – 1:00:36

critical thought process that we

1:00:36 – 1:00:37

need to engage in personally,

1:00:38 – 1:00:39

because if we don't, we're going

1:00:39 – 1:00:40

to lose it.

1:00:42 – 1:00:44

Your comments are making me

1:00:44 – 1:00:45

think about one of the tools

1:00:45 – 1:00:45

we're building.

1:00:46 – 1:00:47

So for FarmFlow, we have the

1:00:47 – 1:00:48

whiteboard tool.

1:00:48 – 1:00:49

We call that Grow.

1:00:49 – 1:00:51

We have a tool that we call

1:00:51 – 1:00:52

Track, which is just managing

1:00:52 – 1:00:53

paperwork. So your bill of

1:00:53 – 1:00:54

lading is your grain inventory.

1:00:54 – 1:00:56

We use AI lightly there, just

1:00:56 – 1:00:57

image recognition.

1:00:57 – 1:00:58

You can take a photo of your

1:00:58 – 1:01:00

grain contract, extracts it.

1:01:00 – 1:01:01

can generate bill waiting

1:01:01 – 1:01:02

numbers. Pretty simple stuff.

1:01:03 – 1:01:05

The thing I'm excited about

1:01:05 – 1:01:07

is the feature we're calling

1:01:07 – 1:01:08

FarmFlow Guide,

1:01:09 – 1:01:10

which is basically a

1:01:11 – 1:01:12

SOP

1:01:12 – 1:01:14

generator, standard operating

1:01:14 – 1:01:15

procedure generator, or best

1:01:15 – 1:01:16

practice generator.

1:01:17 – 1:01:18

If I was to say, John, I want

1:01:18 – 1:01:19

you to go

1:01:19 – 1:01:20

plant corn,

1:01:21 – 1:01:24

and you would have a particular

1:01:24 – 1:01:25

way of planting corn at your

1:01:25 – 1:01:26

farm. Maybe it's based on your

1:01:26 – 1:01:27

soil type, maybe it's based on

1:01:27 – 1:01:28

the equipment you have,

1:01:29 – 1:01:30

maybe the variety you have.

1:01:30 – 1:01:31

And the way that I would plant

1:01:31 – 1:01:32

corn at my farm is going to be

1:01:32 – 1:01:33

wildly different, even if we

1:01:33 – 1:01:34

have the same piece of

1:01:34 – 1:01:35

equipment.

1:01:35 – 1:01:37

And every farm is unique in this

1:01:37 – 1:01:38

way.

1:01:38 – 1:01:40

And so what we want to do is

1:01:40 – 1:01:42

build a tool that partners with

1:01:42 – 1:01:43

the farmer to capture the

1:01:43 – 1:01:44

processes,

1:01:44 – 1:01:46

to document the on -farm best

1:01:46 – 1:01:47

practices.

1:01:47 – 1:01:49

And the purpose behind this is

1:01:49 – 1:01:50

to do a couple of things.

1:01:52 – 1:01:54

to capture and help retain the

1:01:54 – 1:01:57

knowledge of wisdom from

1:01:57 – 1:01:58

retiring farmers.

1:01:59 – 1:02:00

Farmers are like great

1:02:00 – 1:02:01

innovators, great thinkers.

1:02:02 – 1:02:03

They're not necessarily great

1:02:03 – 1:02:04

teachers, or they don't think

1:02:04 – 1:02:05

about themselves as great

1:02:06 – 1:02:07

teachers.

1:02:07 – 1:02:08

And so this tool can be kind of

1:02:08 – 1:02:09

the

1:02:10 – 1:02:11

partner to help the farmer

1:02:11 – 1:02:12

become a teacher,

1:02:12 – 1:02:14

to teach next generation.

1:02:14 – 1:02:16

It can help a farmer scale.

1:02:17 – 1:02:19

So to train employees,

1:02:20 – 1:02:20

interns,

1:02:20 – 1:02:21

whatever it might be, that you

1:02:21 – 1:02:22

can take your practices

1:02:23 – 1:02:26

and give these SOPs to your

1:02:26 – 1:02:28

employees or your team to help

1:02:28 – 1:02:29

them learn and they can interact

1:02:29 – 1:02:30

with it rather than just a

1:02:30 – 1:02:31

static document.

1:02:31 – 1:02:32

It's an SOP that you can

1:02:32 – 1:02:33

interact with.

1:02:33 – 1:02:34

But the part that I'm most

1:02:34 – 1:02:36

excited about for FarmFlow Guide

1:02:36 – 1:02:37

is that I think it could

1:02:37 – 1:02:38

actually be an unlock for

1:02:38 – 1:02:39

adoption.

1:02:39 – 1:02:40

So more

1:02:41 – 1:02:43

than an incentive payment,

1:02:43 – 1:02:45

more than an inspiring

1:02:46 – 1:02:47

video that you might watch,

1:02:47 – 1:02:49

if a farmer is going to adopt a

1:02:49 – 1:02:50

new practice, it's probably

1:02:50 – 1:02:52

because they see their neighbor

1:02:52 – 1:02:53

doing it. They see somebody that

1:02:53 – 1:02:55

they know or have contacts

1:02:55 – 1:02:56

that's similar to theirs doing

1:02:56 – 1:02:57

something different.

1:02:58 – 1:02:59

Farmers trust farmers.

1:02:59 – 1:03:01

When I think about adoption of

1:03:01 – 1:03:03

organic practices or gendered

1:03:03 – 1:03:05

practices on my own farm, it's

1:03:05 – 1:03:06

anything that we've adopted or

1:03:07 – 1:03:08

brought on to our farm.

1:03:08 – 1:03:09

It's probably because I've seen

1:03:09 – 1:03:09

another farmer do it

1:03:09 – 1:03:10

successfully,

1:03:11 – 1:03:12

or they've shared an idea or

1:03:12 – 1:03:13

concept.

1:03:13 – 1:03:14

Maybe it was leaning over the

1:03:14 – 1:03:15

back of a pickup truck at a

1:03:15 – 1:03:17

field day and we were swapping

1:03:17 – 1:03:18

ideas on cultivators.

1:03:19 – 1:03:22

But we want to use an AI model

1:03:22 – 1:03:24

that's trained to think like

1:03:24 – 1:03:25

a coach,

1:03:25 – 1:03:27

to think like an extension

1:03:27 – 1:03:28

educator,

1:03:28 – 1:03:30

and help document the

1:03:30 – 1:03:31

practices that are on that

1:03:31 – 1:03:33

individual's farm, and then

1:03:33 – 1:03:34

create these SOPs that can

1:03:34 – 1:03:36

either be used internally or

1:03:36 – 1:03:37

externally to

1:03:39 – 1:03:41

adapt and adopt new practices.

1:03:41 – 1:03:42

So we're calling this thing Farm

1:03:42 – 1:03:44

Flow Guide just in the early,

1:03:44 – 1:03:45

early stages. So we're working

1:03:45 – 1:03:46

with the same set of farmers.

1:03:48 – 1:03:49

We're really wrestling right now

1:03:49 – 1:03:50

with the

1:03:50 – 1:03:51

data protection.

1:03:52 – 1:03:53

That's an unanswered question

1:03:53 – 1:03:54

for me.

1:03:54 – 1:03:56

We're spending time talking with

1:03:56 – 1:03:56

experts in the field.

1:03:56 – 1:03:57

They kind of think about how we

1:03:57 – 1:03:58

protect

1:03:58 – 1:04:00

this data because we

1:04:01 – 1:04:03

don't want to just train a model

1:04:03 – 1:04:04

and then get that extracted,

1:04:05 – 1:04:06

that value in that process.

1:04:06 – 1:04:08

that's intimate and important,

1:04:08 – 1:04:09

and really the intellectual

1:04:09 – 1:04:10

property of that individual

1:04:10 – 1:04:12

farmer. So that's something that

1:04:12 – 1:04:13

I don't have good answers to

1:04:13 – 1:04:14

yet.

1:04:14 – 1:04:15

But we're thinking about that

1:04:15 – 1:04:18

SOP guide tool

1:04:18 – 1:04:20

as kind of an unlock

1:04:20 – 1:04:23

for farm flow and for adopting

1:04:23 – 1:04:25

new practices in our context.

1:04:25 – 1:04:26

So that's, to

1:04:28 – 1:04:30

your point about the tool and

1:04:30 – 1:04:31

the user of the tool, I think

1:04:31 – 1:04:32

this is like

1:04:32 – 1:04:34

the next level for us as we

1:04:34 – 1:04:35

think about AI.

1:04:37 – 1:04:38

Well,

1:04:38 – 1:04:39

it's certainly a very rapidly

1:04:39 – 1:04:40

evolving and rapidly iterating

1:04:40 – 1:04:41

space.

1:04:41 – 1:04:44

And simply entering into the AI

1:04:44 – 1:04:45

space means that you've entered

1:04:45 – 1:04:47

into a race and you'd better be

1:04:47 – 1:04:48

prepared to move very rapidly.

1:04:48 – 1:04:49

Because

1:04:49 – 1:04:52

the capabilities of these tools

1:04:53 – 1:04:53

are

1:04:54 – 1:04:55

already today,

1:04:57 – 1:04:58

it's

1:05:02 – 1:05:04

There's not a good way for me to

1:05:04 – 1:05:05

quantify it for a general

1:05:05 – 1:05:07

audience, a farming audience, of

1:05:07 – 1:05:09

how rapidly improving these

1:05:09 – 1:05:11

tools are in their mathematical

1:05:11 – 1:05:12

capabilities, in their logic

1:05:12 – 1:05:13

capabilities.

1:05:13 – 1:05:16

but there is

1:05:16 – 1:05:18

such rapid improvement.

1:05:18 – 1:05:19

What I can say is

1:05:20 – 1:05:22

you cannot look at a tool

1:05:24 – 1:05:27

such as, let's say, Gemini or

1:05:27 – 1:05:30

ChatGPT or Grok, you cannot look

1:05:30 – 1:05:31

at a tool six months ago and

1:05:31 – 1:05:32

have tested it,

1:05:33 – 1:05:34

say,

1:05:34 – 1:05:36

okay, I'm going to upload this

1:05:36 – 1:05:37

type of data into it and I'm

1:05:37 – 1:05:39

going to ask it to help me

1:05:39 – 1:05:40

analyze it and come up with a

1:05:40 – 1:05:41

set of recommendations.

1:05:42 – 1:05:43

You cannot evaluate that tool

1:05:43 – 1:05:45

six months ago and say, oh, that

1:05:45 – 1:05:46

failed to give me anything

1:05:46 – 1:05:47

useful and therefore it's

1:05:47 – 1:05:49

worthless and I'm not going to

1:05:49 – 1:05:50

bother using it again.

1:05:50 – 1:05:51

If you test it again today,

1:05:52 – 1:05:55

it is going to be substantially

1:05:55 – 1:05:57

better than it was six months

1:05:57 – 1:05:57

ago.

1:05:57 – 1:05:58

It's not not a little bit

1:05:58 – 1:05:59

better,

1:05:59 – 1:06:01

not five or 10 percent better.

1:06:01 – 1:06:03

I'm talking dramatically better.

1:06:03 – 1:06:04

And so the whole space is

1:06:04 – 1:06:06

iterating, evolving so rapidly

1:06:06 – 1:06:08

that even if you used it six

1:06:08 – 1:06:09

months ago or two months ago and

1:06:09 – 1:06:11

you said, oh, this thing wasn't

1:06:11 – 1:06:11

all that great,

1:06:12 – 1:06:13

you need to stay on top of it.

1:06:14 – 1:06:15

You need to look at it,

1:06:16 – 1:06:18

constantly be evaluating how you

1:06:18 – 1:06:19

are using it,

1:06:19 – 1:06:20

how to get the best results out

1:06:20 – 1:06:24

of it, because it is growing at

1:06:24 – 1:06:25

a rapid pace.

1:06:30 – 1:06:31

So that

1:06:32 – 1:06:33

means that all the users of

1:06:33 – 1:06:34

FieldLark are

1:06:34 – 1:06:36

going to be continually

1:06:36 – 1:06:37

delighted with all the new

1:06:37 – 1:06:38

things that they get out of it

1:06:38 – 1:06:39

from a month -to -month basis,

1:06:40 – 1:06:41

and from FarmFlow perhaps as

1:06:41 – 1:06:42

well.

1:06:42 – 1:06:43

Yeah,

1:06:44 – 1:06:46

the trick

1:06:48 – 1:06:49

I think is to stay grounded in

1:06:49 – 1:06:50

it too.

1:06:54 – 1:06:56

Be aware to interact and then to

1:06:56 – 1:06:56

stay

1:06:58 – 1:06:59

connected to the bigger version.

1:07:00 – 1:07:00

Absolutely.

1:07:00 – 1:07:01

The why.

1:07:02 – 1:07:04

And be very deliberate about

1:07:04 – 1:07:05

what you use it for and what you

1:07:05 – 1:07:06

don't use it for.

1:07:08 – 1:07:09

Wonderful. Well, Matthew, this

1:07:09 – 1:07:10

has been an awesome

1:07:10 – 1:07:11

conversation. We haven't gotten

1:07:11 – 1:07:12

to talk about the farming yet.

1:07:14 – 1:07:17

But I want to be conscious of

1:07:17 – 1:07:18

your time and of all of our

1:07:18 – 1:07:18

listeners' time.

1:07:18 – 1:07:19

So thank you for being here.

1:07:19 – 1:07:20

Thank you for the work that

1:07:20 – 1:07:21

you're doing. And let's stay

1:07:21 – 1:07:22

connected.

1:07:22 – 1:07:23

Thank you, John.

1:07:23 – 1:07:24

I hope you have a good winter.

1:07:24 – 1:07:25

Stay warm.

1:07:25 – 1:07:26

Thank you.

1:07:26 – 1:07:26

Bye.