Articles, Blog

WGS17 Sessions: The Evolution of Food

November 5, 2019

Couldn’t ask for a better introduction than being called the head nerd farmer. Okay, I lead a group at the MIT Media Lab. If you don’t know the media lab,
it’s a place on campus. It’s kind of a think tank,
and so they develop everything from new cars and new toys
to synthetic neurobiology, to bio mechatronics and even food. So, when I approached this in my work, I started with this picture which I think follows upgrade
on the last presentation. If I said to you food crisis, It’s likely that every one
of you in this room would have a different thing
in your mind. You might be thinking
not enough food, You might be thinking too much food, agricultural runoff
creating algal blooms, not enough water, food insecurity
GMO is saving the world, GMO is the blight of the world, farmers’ lands taken away from them, that land was taken
from a rainforest, attenuation of nutrition, urbanization. It goes and on and I think
the current message of food crisis is incredibly disempowering, because by the time
you hear all that stuff you’re like “Oh man!
someone should fix this big problem.” So the derivative as it goes
to the UAE in my opinion. so the UAE imports
eighty-five percent of its own food, probably not a huge shock
to those that live here, at the cost of fifteen percent GDP,
which is quite significant. So you can imagine
the costs of imports are rising because of climate change. Food supply
chains are getting more insecure. So this could be a scary thing, but it could also be the biggest
opportunity of this generation to take back
part of that eighty-five percent and to take back
part of that fifteen percent GDP. So these are my what-ifs for my work
that I’m trying to make happen. What if there was climate democracy? And what I mean is this map is the map of the most beneficial
climates in the world for agriculture in green, the least in red. Climates are changing and moving
and Californian farmers are now Mexican farmers. China is the largest
landholder in Brazil, and a lot of food from the UAE
is coming in from Africa. So we are currently slaved to climate. Could we start producing
beneficial climates anywhere? The next question super serious; If you’ve ever seen this part
of Willy Wonka, they just zapped the candy bar, they turned that candy bar into data. It flows through the air
and it goes to the other side, and becomes a candy bar again. The world today is described
more by data than ever before. When you have Pokémon Go
becoming a billion-dollar company in a week with more users
than Twitter and Tinder combined from a silly game
that you flick on your phone, you need to know that we can
create enough data about things that now maybe
this is possible with food. We ship data not food. The last one, I started
this work post-Fukushima. I went to Japan,
I landed that day, and the headline read
“Japanese agriculture” “has no youth, no money,
no land, and no future.” It isn’t only in Japan. In the United States,
2% of us are farmers, the average age is 58. As you keep going, eighty percent
of the African young population doesn’t want to be farmers. They’ve
left, they’re going into the cities. They’re looking for new opportunities. In India, more farmer suicides
in the last three years than the previous ten. People are leaving farming
in droves all around the world. So who are that next generation? And we’ll talk about that. So I look at movements,
movements that I’ve seen it in my life have been adjacent to or a part of. So coder movement,
access to hardware, Linux HTML, the building of the internet, maker movement were kind
of maybe towards the tail end of, network movement and then
the last one, food movement. In the last two years,
six out of the 30 labs of the Media Lab are doing biology. This is a phenomenon
that has never been the case since the inception
of the lab in the 80s. So we are starting to use digital tools with a new fluency towards biology, that’s exactly like that
60s 70s era of computing. So this is part of my work,
one is called production platforms. You’ve probably seen
things like this before. This is one of the first labs I built, and we can talk about the technology. This was five or six years ago, but I think the most important thing
was that once a month, we had a harvest and right
in the middle of the media lab, bright white roots,
deep green colors. Is this a new grocery store? Is this
a new school cafeteria experience? Those are big questions
but I know for one thing this is the first time anyone
in the media lab ever ripped the roots off of anything. We get our salad in bags. This growing is not a part of our life. So this moment inspired roboticists, algorithms experts, data scientists, mechanical electrical engineers
to get involved. So my next lab
is on the facade of the media lab. And if you know about
the genome or genetics, what we’re talking about is the phenome or the phenomena
that surrounds genetics. In here we design climate. Now if that sounds funny
think about if you planted a seed and the same seed around the world, you might end up saying
“that tomato in Tuscany” “on the south slope in the summer time
with the sheep next door” “that’s the best tomato.” It’s really the environment
or the phenomena that surrounds it that causes an expression of genetics. So inside of this lab
we’ve been designing climate to produce flavor. You can imagine that climate recipe
as that blue line. So it’s trying to follow 1982 Napa. We’re going to produce
the 100-point Cabernet. All those zigzaggy lines
are all the sensors that we’re following to try to conform to that climate. We correlate those recipes
to each plant. So in this each plant has an ID, all plants next to it are kind
of arranged as friends, and what we’ve essentially created is the Facebook of plants. Each user unique,
each user’s needs unique. Starting to see
about combinatory effects between plants that no one’s
ever known before. This is the lab a few months ago. We’ve been growing all kinds of things. A lot of people like to talk
about aeroponics. Aeroponics was invented in the 80s for a mere space station. It’s a method of reducing
water consumption. It’s just another kind of irrigation. So inside of that box,
there’s no soil, there’s no standing water, only mist. And that mist will grow things
three to five times as fast. We’ve also been growing heirloom
ancient and rare genetic. These tomatoes
came from a rare seed bank. So when you think about
the tomatoes you get to eat, there are about five
cultivars of tomatoes that you get to put in your mouth,
and they were never picked for flavor, they were never picked for anything
other than supply chain, ability to withstand supply chain. So these tomatoes
hadn’t been grown in 150 years. They’re supposedly the best
sauce tomatoes in the world. Of course we don’t know how to cook
so we’re not sure of that. We’re also growing cotton.
One of my sponsors is the world’s largest
cotton manufacturer. The water table in India
is dropping feet per year. The cotton quality and consistency
is threatened globally. So the question is can you have
cotton growing in a factory right next to your
ginning spinning milling product making, out the other side? For the first time
in history, agriculture being 365 days consistent would fundamentally
change the economics. We knew these machines
were going to be expensive and we needed to curate a community, so we created what we call
the personal food computer. Think of this
like the first 3d printers, or the first like… people that worked on computers
in their garages during that kind
of home brew computer club days. These are weird,
they aren’t meant to feed people, they’re meant to create knowledge
and they’re meant to create curiosity. We drop these food computers
off in classrooms. First the student sees
the digital climate, sees the sensors, the actuators, loads a recipe that some other student anywhere in the world
has grown with before. When they load that recipe,
the climate begins. They plant the plant,
but they get curious. They say “why does the plant need CO2?” “Maybe it doesn’t need CO2,
crank down the CO2.” Plant dies, kid learns a lesson
about climate science. At the end of it they harvest
physically and digitally, and when they harvest,
they get a derivative recipe. They’ve now added
to the commons of knowledge about what happens to this crop
under different climates They download the data,
they do Google science fair, or whatever else they do with the data, but importantly it’s captured. We’ve had students do climate studies. So on the left side you’re seeing a climate
that was 50 years old. They used farmers
almanac data from 50 years ago to reanimate a climate of the past. They then used data gathered now
from GIS and other sources to animate the climate of the present, then the students predicted
the climate of the future based on some of the predictions
we heard today. It’s a lot more interesting
for a young person when you say climate change
and they look up and they don’t see anything to when they have climates
at their disposal to design in their classrooms to understand
the effects of what’s going to happen. We released the second version
of the personal food computer at the White House a few weeks ago. I’m proud to announce
on this stage that right now we’re deploying them
with the World Food Programme in the UN in Amman, in a Syrian refugee camp to start innovation around
growing food there this week. So everything’s moving fast. This is one of my new labs
that I just built at MIT. We are putting an investment
into this space. We’re scaling up our work
into shipping containers and beyond. So if you want to be involved in that,
let me know. The second part of the work,
so that was all hardware. Now it’s all data. So I told you about the phenome, and that’s an incredibly complex idea, because think of all the variables. You know the genome,
that’s kid stuff. We at least can figure that out
because it’s easier, but if you’re trying to figure out
the comminatory effects of 60 variables on an organism, you’re going to need
some massive computing power. So inside of all the food computers, we collect a lot of the
standard data you would expect; things like temperature
and light and humidity and so on. But we also collect mineral data. So this is using spectrophotometer, so when people say
“but it’s the terroir.” “it’s the terroir
that makes it taste good.” Terroir means minerals
and carbon and water and Oxygen and bacteria and microbes,
it’s not magic. And so in here we look
through all those minerals that are being applied to the soil
and what their effects are. We also take the root zone
and culture it. So we just like stick a swab in there,
put it on a petri dish, and then we sequence it and we say
what was alive at that root zone. So when people talk about
microbiome right now, what you’re seeing
is the microbiome of a plant. In the bottom set
of the petri dishes is the roots, in the top is the leaves. The leaves are sterile, the roots
are full of a vibrant ecosystem of microbes and bacteria that today
are not even fully understood. So we’re collecting
all of this data in concert. The last one when the plant is finished, we do analytical chemistry. We want to know what’s in it,
when people say it’s good, why is it good? What is flavor? How do you quantify it?
In this we use gas chromatography and mass spec to get down
to the compound level. So if we want to know what good is, now we can link that
to what bacteria made it good. We use computer vision because
it’s non-invasive and it’s cheap. We would love to replace
all the sensors that we use, and someday we will with new advances
in computer vision. We tie all of that together. So three and a half million
data points per plant, per grow fed into a recipe
that goes into a robot. So this robot plans, acts,
creates climates, learns, and then does it again over and over. We take all that data forward
to do machine learning. So what you’re looking at here
is Bayesian optimization and Gauss’s processing.
What that means is… See these little slices, each slice represents
a two-dimensional graph, so you can test two variables. But when you stack them
against each other, you can test a combination of variables
and create a heat map. So when you’re looking
at the little red dots, this is telling us that
out of 3 times 2 so 6 plus another axis, that amount of combinations
is giving us a hot spot in some place
that we didn’t know before. So it’s fundamentally new knowledge only possible through massive data sets, which we’re just beginning to build, and new tools that everybody’s
working on, AI and machine learning. These things of course exist
in the world. Probably Japan’s the world leader with something around
the neighborhood of like, I think last I heard 800 hectares
of what they call plant factories. So this plant factory grows
a million heads of lettuce a week, or so I’m told. And you can see
it comes out branded Toshiba, and people still eat it. In Japan, the condition
is genuine fear of food. We’ve been fed things before
that killed our children, we’ve been fed things before
that had arsenic and lead and possibly radiation.
We want clean food. So this is why you’re going to see this
fetch a 3 times premium in the market. Of course it doesn’t
just end at lettuce. So this is a project
that grows something very special. Looks kind of the same,
looks like things you’ve seen before, but is actually growing Ebola vaccine. So this is tobacco plants
in a very specific climate stressed in a very specific way to produce a resistant protein. They harvest that protein
off the plants, they manufacture it into vaccines. This is one of the reasons
we got ahead of Ebola. This is a project with the department
of Defense and Texas A&M. So everything from pharmaceutical
natural derivative, cosmetic, nutraceutical
all the way down to lettuce and herbs. But then reality hits us in the face. So… First one is the first vertical farm
in the world on Vancouver that went bankrupt in 2013. The next one is FarmedHere,
went bankrupt 3 weeks ago. This is you know millions and millions
of dollars of investment, and I think this points to the stage
in current technology development that we’re in. It’s early.
It’s like cars before Ford, when cars were super weird
and went too fast or too slow, were all designed by one company. They weren’t integrated
like cars became later. If you were designing a wheel,
you were also designing the motor and the seats and the body, while you probably aren’t good
at all of those things. This is where we are right now; a lot of proprietary intellectually
and intellectual property being sold to very high questions of what
the value of that property is. And so I’m taking a cue
from the market and my work. As I span the big tech leaders, one of which will talk today, see Elon Musk and he open-sources
the patents of Tesla. Why? Why would Elon give away the thing that he spent a ton of money developing to the general public?
Makes no sense. He’s a crazy person. He’s not crazy,
he needs more charging stations, and if he needs more charging stations,
he can’t also build those. He needs smart roads
that work with smart cars. He’s not going to be able
to build those. So he needs people
to understand what he’s doing so that they can supply infrastructure
to work with him. Apple, the most proprietary
company of all time going open source
with their app developer language. Facebook, the biggest asset
they have right now their AI called Torchnet Torchnet, TensorFlow
going open source. Human Genome Project, one of the
most influential science projects of this generation was open source, because it had to be,
it has to hit the network, you have to test the knowledge, and you need a community
of participators with you along for a journey
of changing the world, and that’s what we need to do in food. So all the work from our lab
is open source, hardware, software, and data, we share it out and we look
and see what happens. And these crazy people around the world have downloaded our videos
for the small food computer because it’s reasonable
to build yourself, and they started building it
all over the world. We then started a forum because they needed to get together
and talk to each other. So we started a little
Reddit style forum. Please if you’re interested in this and you feel like maybe
you want to contribute but you don’t know how or where, there’s whole bunch of you in our forum that are asking
each other questions, so please login,
become a farmer, and contribute whatever it is you have. This has spurred a kind
of maker movement of food. So this is a thing we did
with National Geographic where we collected students
from all over the United States. They built their own
personal food computers in two days,
they took them to their schools, and they started home brew
food computer clubs, exactly the same thing that was going on when you saw these guys like Gates
playing around in their garages. And of course,
the network is really important, so working on with projects
like Museum of the Future, Auto Farm, I hope you guys
go check it out. But the question is how can
our tech, your tech, all of the different people
working on this somehow benefit from the network? So that’s what we’ve built now. This is the first time
I’ve really started to talk about what we’ve been planning. so we have the research
and if you look into our work, you’ll find that last month,
secretary Kerry former secretary Kerry
announced a partnership between us and Vietnam
with Can Tho university. We’ve expanded with partnerships
in Chiba university in Japan. We’re talking about
Museum of the Future now here. So that’s on research, and then you need something
to protect the commons. So we’ve created a non-profit because you don’t ever want
that data to be sold, you don’t ever want
the hardware to be sold, and you don’t ever want
the software to be sold. If HTML and Linux were proprietary, the internet would have never existed. So we’ve created a fully non-profit
device for putting all that in. And now it’s spurred
this commercialization layer. So with some of our partners
like Food and Future that’s here, and going to talk with other spin-offs
that have come out of my group. They’re now taking open-source tech, raising VC funds
which everybody wants to put in to change the world of food. So this is exciting and new. Our work has now expanded
to 20 countries in six continents without a whole lot of major funding. This is just people that were excited,
that were interested, that downloaded our code, that
built our funny-looking little machines, and started putting data up
about what they were doing. This is the world today; a complex system. We’ve got any country
with a color its food insecure. That’s 80 percent of the world
is food insecure. Any country that’s both red
and then purple is a country that is providing food
for another country. So most of them are food insecure
and providing food to other countries. Then you’ve got the complex
distribution lines on top of it. Planes, trains, and automobiles, our inheritance
from the industrial revolution. This is what we’re dealing with today; a very big very global farm that was very good
for the call in the 70s which was more cheap food. Now we’re asking for better, more
sustainable, more nutritious food, and this is the future. We will start bringing online
digital farms that are sending climate files to other farms to animate Mexico in Abu Dhabi or to animate Burgundy in India. It’s already started to happen.
I invite all of you to participate in whatever way you want, supporting the education,
building the machines, creating commercial ventures
off of open systems. The future of food is not
about everybody camping off which is what’s happening now.
“no, GMO is bad,” “no, it’s good, no,
organic is good, no, it’s bad.” It’s not about that. It’s about networking
the next 1 billion farmers with common and open tools to simply ask the question
“what if?” Thank you.

No Comments

Leave a Reply