Latest / The Future Herd / 37: Improvisation and Intuition Are the Requirements for AI in Agriculture

Transcript

37: Improvisation and Intuition Are the Requirements for AI in Agriculture · The Future Herd

0:00
Jesse Hirsh: Hi, I'm Jesse Hirsch.
0:03
Welcome to the Future Herd.
0:18
This is an episode that in my mind, really demonstrates the
0:22
potential of the future herd.
0:24
On the one hand, our mandate is to look at leadership within the AgriFood
0:30
sector, but on the other hand, we're trying to anticipate the future, uh,
0:35
in a manner that allows for leadership, that allows for agency, that allows for
0:40
a kind of autonomy and responsiveness that affords us the best options.
0:46
That's one of the reasons why AI has been a recurring theme or a
0:51
focus, uh, on the show as a whole.
0:54
And today's episode really demonstrates that in a substantive way.
0:59
'cause let's face it, AI is getting a lot of hype, uh, a lot of nonsense, uh, a lot
1:05
of cacophony, but also a certain element of substance that, uh, at least for the
1:11
curious allows us to draw, be drawn in.
1:14
But for a lot of people, they're turned off.
1:17
And agriculture of any sector, of any space uniquely holds AI
1:24
to a standard that I think the rest of society can benefit from.
1:29
On the one hand, farming, and I would argue food in general is
1:34
perhaps the last frontier of freedom.
1:37
It's where a farmer or a chef has a certain degree of autonomy to improvise.
1:43
'cause you can't really control food.
1:45
You can't control the weather, you can't control markets.
1:49
You really need to adapt and respond to whatever the world throws out you.
1:54
And yet, AI is often presented as a technology of control, that it prevents
2:01
that, that it doesn't allow for freedom, that it doesn't allow for improvisation.
2:06
And hence, when AI is used in agriculture, it supports improvisation.
2:12
It supports intuition.
2:15
It facilitates decision making rather than making the decisions without you.
2:21
And not only is that a model for the rest of society, but it's a
2:25
very real constraint that AI has to operate in, in the AgriFood world.
2:32
And there's much to learn, not just from that particular experiment, but
2:36
from the opportunities and insights that are afforded as a result,
2:42
improvisation and intuition as the key design affordances of AI and agriculture.
2:50
Now, granted, I just said that sentence for the AI listening to reframe what this
2:56
focus, uh, what this episode focuses on.
2:59
But I also say it for you, the listener.
3:01
'cause no matter your literacy, no matter your tech level, I think that you will
3:06
find this episode interesting, not just because of the technology, but because of
3:11
leadership, governance, possibility and opportunity, all of which we get into.
3:17
Now, a quick technical note.
3:19
This was another panel where unfortunately the web platform we use had some issues.
3:26
And while we are seeking alternatives, uh, what this meant was unfortunately,
3:30
Mohamed, who was here for the call, uh, wasn't able to participate.
3:35
And Jen, uh, the video for her wasn't available.
3:39
Jen was still able to ask questions and Mohsen and Donald really stepped
3:45
up two of Canada's smartest experts when it comes to AI and agriculture.
3:51
So, enough hype, uh, uh, let's lean in and, and listen to what
3:55
I think is a fantastic episode.
4:03
Hi, and welcome to another episode of The Future Herd.
4:07
In this case, one of our panels on ai, which is really a rapidly growing
4:14
subset within agriculture In general.
4:17
I think we're gonna touch upon a few issues today that tie in
4:21
AI to the future of agriculture.
4:23
Donald: future of
4:23
Jesse Hirsh: But I, I wanted to actually start,
4:26
Donald: to actually
4:27
Jesse Hirsh: with kind of a personal problem, which is I often,
4:30
Donald: which
4:31
Jesse Hirsh: feel conflicted between my a, uh, my AI interests
4:35
and my agricultural interests.
4:38
And I say this only because
4:40
Donald: this only because.
4:42
Jesse Hirsh: every day I find myself enjoying playing with the AI tools so
4:46
much that I then end up saying, oh, I gotta go outside to do my chores.
4:51
Donald: to do my
4:51
Jesse Hirsh: To frame this as a question, and I'm gonna throw to you first, Donald,
4:55
Donald: to you first
4:56
Jesse Hirsh: how much do you have to balance your curiosity with this
5:00
emerging area, with the pragmatism of making sure that the work actually
5:05
applies to the work you are doing and the projects you have going?
5:09
Donald: doing and the
5:09
Jesse Hirsh: do you balance the blue sky research with the, the building and the
5:13
deployment that's part of the process.
5:15
Donald: we ran into that head on this week.
5:18
You know, I, I'm my team about, um, X 4 0 2 and how agents are gonna be
5:26
identified and, and pay for, um, MCP and, and, uh, API access and data access.
5:33
And I see that as a big part of the agricultural implementation.
5:37
And simultaneous to that, we need a strategy to get our first 10 farms like
5:43
loaded in, and we need a, and we need to be delivering them intelligence in
5:47
a week in which the temperature has soared to 35 degrees and we wash the
5:51
soil moisture, um, rapidly decreased.
5:55
And on top of that, Jesse, I mean, my role encompasses a lot of, of non-tech stuff.
6:00
Like I have to administer the advanced payments programme.
6:03
I have to lead the administration of the On-Farm Climate Action Fund.
6:06
I have to engage with my provincial government and federal government.
6:09
So, um, the thought leadership happens in the margins.
6:15
It happens on the weekend and in the evenings.
6:18
And, um, and then the day to day, uh, it, it, it, it happens very
6:26
appropriately along the lines of what our, our requirements are for
6:30
our funding agreements and, and our farmers and our perspective users.
6:33
And so it's really just one slack channel is blue sky.
6:37
One Slack channel is where we lashing a gateway to tomorrow.
6:41
You know, one Slack channel is what, what should this app look like?
6:45
And, uh, and those are all really fun spaces.
6:48
And it's really as simple as that.
6:50
You know, slack is becoming a huge part of the enterprise integration and we're,
6:55
I'm, I'm an early adopter of Slack and, and I think clear communication channels,
7:00
even if the same people are in them.
7:02
That's how I manage it right now.
7:04
Jesse Hirsh: Right on.
7:04
Now, most in this question is more core to your kind of professional
7:09
activities as a whole in terms of how do you balance the long-term research
7:13
with the short-term research, which
7:15
Donald: short
7:16
Jesse Hirsh: with research methodologies and the tools that are there.
7:19
Donald: the
7:19
Jesse Hirsh: So I'm curious, how do you balance all this, given how exciting both
7:24
the technology and the larger research that you're currently engaged in?
7:27
Donald: currently?
7:28
Mohsen YN: Yeah,
7:29
Donald: Yeah, that's a very great question.
7:31
Uh,
7:31
Mohsen YN: Jess,
7:31
Donald: that,
7:32
Mohsen YN: is, so
7:34
Donald: so
7:34
Mohsen YN: my, in
7:35
Donald: my, in my programme,
7:36
Mohsen YN: I'm first
7:37
Donald: first plant leader
7:38
Mohsen YN: computational biologist.
7:39
So in the
7:40
Donald: in the plant
7:41
Mohsen YN: I have a,
7:42
Donald: I have a,
7:42
Mohsen YN: in every plant breeding programme, we have a
7:45
Donald: we.
7:45
Mohsen YN: long term goal of releasing varieties based on different criteria
7:50
that we're selecting in the field.
7:51
And then, you know, collecting so long, like, you know, uh,
7:55
all different from all different traits and everything every year.
7:59
So.
8:00
Basically how I'm doing is I'm signing as a sub manager to
8:07
different areas in my programme.
8:09
Let's say I have right now three active and full-time technicians,
8:13
research technicians in my lab.
8:15
And I have one, uh, lead technician, uh, Lindsay Schram.
8:19
So Lindsay is responsible for maintaining day-to-day activities that happening
8:24
in the field, and also for making crosses and all those kind of things.
8:28
And also managing technicians.
8:30
It doesn't mean that I'm going to give, like, you know, be far very far
8:35
away from the fields and everything, but at least I have a chance to just
8:39
meet with Lindsay every week and see if there is, if everything goes well
8:44
or not, or how can I help them, right?
8:47
So that's, that's, that's, that's basically for, um, managing the
8:51
fields and also plant breeding.
8:53
And for long-term goal, like, you know, for all these kind of things that makes
8:57
need to be happening in the field, we need to have long-term funding.
9:02
So that's my main responsibilities.
9:04
And, uh, every day my, every like day-to-day task is to write, like,
9:09
you know, more research grant, talk to different industries, talk to, uh,
9:14
government buddies, like farmers, try to understand what are their needs,
9:19
what are their demands in the future, and then how can I collaborate with
9:24
them and how can I sustain my programme smoothly and run my programme smoothly
9:28
without any hiccups on the way.
9:31
So that's, that's for, uh, plant breeding side for the computational
9:36
biology, all of the computational, but most of the actual computational
9:39
biology things are in my shoulders.
9:41
So I'm trying to, and it comes from my curiosity.
9:44
I love to just, you know, know more about ai, know more about like, you know,
9:48
the company new computational methods, how can I, like, get the best use out
9:52
of the data that comes from the field.
9:54
And all of my technicians are collecting all those kind
9:57
of, uh, like, you know, data.
9:59
So now it's my time and it's my turn to make this data into the meaningful,
10:04
uh, like definition something, right?
10:06
So in this case, I'm trying to, uh, a couple of hours at least per day
10:13
to just read more paper, see how can I, like, you know, contribute
10:17
efficiently to my programme in this case.
10:19
But also I have postdoc, I have students, and I also have very talented volunteers
10:27
and very talented, like undergrad, as soon as they are very passionate.
10:30
And I think that, you know, they are going to be the next generation plant breeder
10:33
in the, in the very like, near future.
10:36
And it's not the actual only traditional plant breeder.
10:39
There are going to be the computational plant breeder.
10:42
So they are, we are all passionate about these new tools, new, uh,
10:47
you know, like, you know, uh, way to collect and decipher the data.
10:52
And yeah, we are all working together.
10:55
In terms of, if I, if I, like, if you ask me that, if I have a very relaxed
11:01
weekend or not, I can't tell you because over the weekend I usually like, you
11:07
know, uh, wanted to keep the pace with the amount of the emails, amount of the
11:11
things that I'm receiving every day.
11:13
And I remember in the last like, you know, LA uh, in this week, like it was Canada
11:17
Day a couple of days ago, so I took some days off and I went to the cottage and
11:23
after I came back I faced with 350 emails.
11:27
So had to just reply, reply all those emails back because I wanted to
11:34
respect everyone who sent me an email.
11:36
And then it took me like 10 hours not to stop just replying back to those emails.
11:42
So yeah, that's a life, that's the work.
11:45
It's exciting, but at the same time, there are so many responsibilities and
11:49
I'm trying to like keep the pace with all those kind of things that happen.
11:53
Jesse Hirsh: Right on.
11:54
And I appreciated both of your answers because it really reinforced why as
11:59
a leadership podcast, as a leadership within an agriculture podcast, we're
12:05
talking about this technology because both of you kind of described how
12:10
leadership characteristics, leadership responsibilities, and leadership vision
12:16
is really a key ingredient in using these technologies quite effectively.
12:22
I bring up this point about leadership because often I feel people who
12:27
don't have a technical background are really intimidated by this technology.
12:32
They, they kind of disqualify themselves and think that they're
12:35
not an expert, they can't use ai.
12:38
So I'd love to have a, a, both of you kind of weigh in on the role of language, how
12:45
even though there is still a math, uh, a science, uh, a, a technical on the level
12:51
of software, uh, underpinning to this entire kind of ecosystem, part of what's
12:57
exciting right now is that you can use language to interface with these systems.
13:02
You can use language to build these systems.
13:05
I think those of us who are using it take it for granted, but I think
13:10
outsiders don't really understand how this is a different way of engaging
13:15
with the research, a different way of, of building applications.
13:19
So I'd love to have both of you, from your perspective, talk about kind how language
13:25
plays a role in what we talk about.
13:27
Ai a big abstract subject, but I think something that'll open up a bunch of
13:32
threads that, that we can pull from there.
13:35
Donald, maybe, do you wanna start us off?
13:38
Donald: Yeah.
13:39
I think, um, you know, I mean language as a code for, you know, if we we're
13:45
signifying, you know, the signs and how we signify them through language
13:50
is, um, a very interesting topic of the human, of the human experience.
13:57
It's one of our most important codes, you know, along with economics
14:01
and, uh, to motion's point, like biology and the genetic code.
14:06
But to, you know.
14:09
To deliver these tools to farmers, it is important to remember that they are
14:14
not interested in the nuts and bolts for the most part of an emerging technology.
14:21
And that's where, not just language, but um, and language and how we talk about
14:27
it is almost as important as the fact that we're using large language models
14:31
if we're going to deliver it to farmers.
14:32
But more importantly is aesthetics, you know, and, and the aesthetic of
14:37
this technology and how it's delivered to farmers and that human design
14:41
element that certainly cannot be, um, outsourced to a large language model
14:48
or, or any type of research loop.
14:50
You know, there is, a last mile piece to this that requires human engagement,
14:58
whether it's in this case, farmers, you know, and it has to be delivered
15:03
to them as a human technology.
15:05
Um, and that's something that we battle with as we, um, you know, prepare
15:10
the, the mobile apps and the, and, you know, the front end development.
15:14
It's how we were so excited in April, may, we had built real artificial intelligence
15:20
for farmers and it worked and it made it easier to make carbon credits and it
15:23
made it easier to, uh, implement variable rate nitrogen and, um, it made it easier
15:29
to maximise the value of supplemental irrigation and the potato crop.
15:32
And we had built all that, and we could do that computation very easily, but
15:37
we were faced with the issue of, okay, aesthetically how are you going to deliver
15:42
this opportunity for greater productivity to to, to Prince Edward Island farmers?
15:47
And that caused us to, to step back and, and consider that.
15:51
And, and the, you know, since the 17th century, like we understand taste
15:57
to be, you know, elegance and, and simplicity and, and, and we have that,
16:03
that nothing about that has changed just because we've broken language
16:07
down into tokens and we can model it and deliver information based on that.
16:13
You know, this is still a human technology.
16:16
It still has our fingerprints and our signature on it, and it has
16:19
to be delivered with, um, the core tenets of, of aesthetics and, and
16:24
of design, human-centered design.
16:26
And that's the most exciting part as we live in this sort of like post a, like
16:33
post ai, but like pre kind of like we still have user experience, we still
16:39
have bespoke ux and we have to design UX that can deliver tremendous amounts of
16:44
computational power to farmers in a way that they simply open it and it works.
17:32
Yeah, look good, feel good, Jen.
17:33
Like not just aesthetics isn't just about looking good, but it's about how, how the,
17:39
how the object really, uh, makes a person feel and, and, and has a huge impact on
17:44
whether they're willing to consider it.
17:46
And so it would be so easy to build something that farmers, um, turn
17:51
their back on when you're talking about this type of emerging technology
17:54
with so much, um, controversy on its impact and its usefulness and,
17:59
uh, so much rhetoric and media.
18:02
And so, you know, it's not just about how it looks, although that's important,
18:06
but how it feels to really use it and, uh, make it a, a welcoming space.
18:12
Um, and, and valuable.
18:14
And I think that really the last two are the ones we think about the most.
18:18
Um, you know, we want, um, our team in involved in the, in
18:24
the, um, implementation of it.
18:25
We saw Palantir in the USDA to the $300 million deal.
18:29
And certainly that's gonna involve Palantir going into the USDA and
18:33
understanding their systems and building things, uh, that fill these crevices
18:38
in their organisations and, and create value added computational opportunities.
18:44
But at the, in the, at the grassroots level with the farmers, there's no
18:48
question that ensuring our first 10 farms have a very strong, positive
18:53
experience, is the most critical part of our project right now.
18:57
Uh, we're working with Prince Edward Islands, some of Prince Edward
19:01
Island's most influential farmers.
19:02
And if we give them a valuable experience over the next, uh, eight weeks, uh,
19:07
I have no doubt that, um, our signup sheet will, uh, will start, um, start,
19:12
uh, really to expand quite quickly.
19:15
Jesse Hirsh: I, I mean there's a lot there Mohsen for you to kind
19:18
of build upon both in terms of the accessibility and the inclusivity.
19:23
But I, I'd also love you, you know, from a research perspective, talk
19:27
about the potential unreliability or the inconsistency of large language
19:32
models and, and how you factor that into the work that you do.
19:37
Mohsen YN: That's a great question.
19:38
So before I talk about like, you know, inconsistency, I just wanted to, uh,
19:43
add one point about the leadership.
19:45
Like, you know, me leadership is, doesn't come from one person alone.
19:51
So it comes from the strength of a very good group.
19:54
And I said, I'm very proud to have a great team around me.
19:58
So they are all bring ideas, energy, honesty, and commitment.
20:03
They the level of the commitments that they have, and they are coming to the lab
20:07
every day and working with the passion.
20:10
it makes my life easy.
20:11
Like my role as a leader just support them, listen to them, and like creating
20:18
a space where everyone can flourish.
20:21
Their idea can be shine, like, you know, and can do their best work.
20:25
So that's something, uh, I, I miss to, I forgot to say about like, you
20:30
know, uh, being a good leadership and how the leadership means to us,
20:34
but in terms of like, you know, when Donald talk about, um, the language.
20:39
And how to talk.
20:41
Like brought up a very good point about like the tokens, how we need to be
20:46
cautious about like, you know, how we are speaking to the AI and also how we
20:51
are, working with like non-AI things.
20:54
But, uh, getting back to your question, Jen, about like, you
20:59
know, farmers, I think, I think it's slightly different than Donald.
21:03
I think like, you know, farmers like AI should not come to,
21:07
to the farm as a black box.
21:10
It should come as a kind of conversation.
21:12
So time that I'm talking to farmers, they don't want something that
21:17
tell them, for instance, spray now or do this without any reason.
21:22
It should says that like, you know, disease risk is high, this part of
21:26
the field because of that, that, that, and then like, you know, based on the
21:30
previous field history, all points in the same direction tells us that like,
21:35
you know, there is a like 90% chance that you need to spray in this area.
21:41
That's something that I think farmer very well like it and better like it.
21:47
And then they are like, tend to more adopting this kind of
21:51
technologies way better than like, do this or do that or just give some
21:54
data and receive the prediction.
21:56
So in my case, I try to involve farmers in different decision like
22:03
we have at University of Guelph.
22:05
We have like, I'm going to develop a sensory panel like inviting
22:10
farmers, growers to actually see the canned beans and ranking them.
22:17
And then I can use this information to build an ai.
22:20
So later whatever AI that I'm building is be, is based on the farmer's
22:26
feeling, farmer's decision, right?
22:28
So this can help them to better understand a better adopt this kind of technology.
22:33
And just, I got surprised week how much farmers are tend to
22:39
adopt these new technologies.
22:41
I sometimes thought thinking that, like, you know, that okay, the, I I
22:47
will have like, you know, more hard time talking to different people, what's the
22:51
benefit of AI and those kind of things.
22:53
But last year I received different calls from farmers that
22:56
they are very much interested.
22:59
this new technology, which like, you know, encouraged me
23:02
to work in this area way better.
23:04
But yeah, so that was something that everyone to talk about it.
23:07
But speaking of uncertainty, definitely it's, and hallucination.
23:11
Hallucination in AI is definitely something that it's makes me like,
23:18
you know, scared about it, uh, but at the same time makes me worry about it.
23:21
But at the same time, I'm trying to like find a good way to
23:26
overcome these kind of challenges.
23:28
For instance, I can tell you one, like there are so many areas that we are
23:32
dealing with, but I don't want it to bring all of them up here, but just give
23:37
you one, some 1, 1, 1, 1, uh, example.
23:41
Let's say that Donald, for instance, in their own area find something special.
23:47
Like he find that, he found that, uh, I dunno, X can cause y, right?
23:56
very interesting things.
23:58
And he decided to bring it up here.
24:02
So he brought it up here and you write a good news.
24:07
So this news will be publicly available over the internet and AI
24:13
agents are going to use this news.
24:16
Then started to think that, okay, how about like, you know, publishing a good
24:21
paper out of this great achievement and he's publishing a paper, so it's the
24:26
same result, but publishing in another source in another way, and AI agent
24:31
also collect those as another evidence.
24:34
Then publishing somewhere else, like so many news, like, you know,
24:39
broadcasters, they are all coming.
24:41
And I said that, okay, we wanted to cover that news.
24:45
So there are so many X caused Y and they are all treated independently
24:53
on the AI agent while they are all coming from one source.
24:57
So if as a researcher, I wanted to ask a question, how, hey,
25:02
ai, what do you think of X?
25:05
So it's a generative, like, you know, is a rag model.
25:08
Retrieval, augmented, generative models goes back to retrieve so
25:13
many things and consider them as an evidence it can search publicly.
25:19
And if you'll find that, okay, X caused y mentioned thousand
25:24
times, so that's something there.
25:26
So it gives you this result.
25:29
you know what X caused Y because of this, this, this, this, this.
25:32
But all of them are coming from one source and it's all in one study.
25:38
So that caused a problem because I know that I'm doing something correct,
25:45
Donald, doing something great, awesome.
25:47
But how about verifying it over the years, right?
25:50
So we need to have such this kind of controller in between,
25:54
or maybe we need to change it.
25:56
So I recently came up with another idea instead of retrieving news,
25:59
how about retrieving keywords?
26:02
So I can add X cause Y as a keyword into an index, like in today's database.
26:08
And consider this as a one time things, right?
26:10
If anyone with another name, with another, like, you know, nature x, cost YI can add
26:17
it as another evidence to this keyword.
26:19
Something like that, right?
26:20
I'm just thinking of it.
26:22
But are so many other ways, and I think that's a kind of need to be corrected,
26:29
need to be, um, actually controlled.
26:33
Or even some of the papers that, uh, in, in my research, I'm sometimes
26:38
searching that, Hey, would you like, you know, uh, can you retrieve some of
26:44
the genes related to something else?
26:46
Like, some part of, in some parts in plan and I'm receiving some of the
26:51
good research paper and sometimes not very good research papers.
26:55
So it need to be a controller there to just sort it out based
27:00
on like the credibility of the authors and also, uh, journals.
27:06
And also sort it out not only based on the impact factor, not only based on that if
27:11
other like publish so many papers or not, there should be something there, right?
27:15
To control all those kind of things.
27:16
So I think that's a very important part and I can't give you the like,
27:22
you know, one answer that hey, yeah, we can fix everything with and we can
27:28
like address all the uncertainties.
27:30
No, still uncertainties there.
27:32
And the more data that we are creating and collecting uncertainty, uncertainty
27:37
will be increased until we are setting back and start filtering, optimising,
27:46
pre-processing all the information and then develop AI based on that.
27:51
Jesse Hirsh: Well, and what I love about that answer.
27:53
Donald: it sounds like Mo wants to de sounds like he wants to design a harness.
27:57
Jesse sounds like he needs a good harness.
27:59
Mo, it sounds like you need a good harness.
28:01
My man.
28:02
Jesse Hirsh: and, and Donald,
28:03
Donald: skills baby.
28:04
Let's go.
28:05
Jesse Hirsh: Donald, that, that helps set up a, a really great question.
28:08
And you know what I'd loved about your answer, Mohsen?
28:11
Is it evoked governance?
28:13
Why policies, why frameworks, why verification is so important in these
28:20
types of systems that we're building?
28:23
I personally wish we had figured this out in the early days of social media.
28:27
'cause every social media platform has a governance platform.
28:31
But doled, you, you correctly evoked the word harness, which
28:36
I suspect for Jen's benefit.
28:38
You know, after our last talk, Jen sent me a note saying, I gotta go
28:42
learn what Hermes is and, and you know, uh, uh, learn a little more
28:46
about the software around these agents.
28:48
And that is, of course, the harness.
28:50
So I'd love for you to kind of do the 1 0 1, what is a harness, but also kind of
28:55
build off what Mosin is saying, which is what are the types of systems, what are
28:59
the types of interfaces that we create?
29:02
Not just to compensate for the potential hallucination, but
29:07
maybe even take advantage of them from a design perspective.
29:13
Donald: Well, keeping our farmers insulated from, the of broad expensive,
29:23
um, frontier AI computation, keeping the data separated from that is definitely
29:30
the part of the development cycle that's, that we're in and is most interesting.
29:34
And a big part of that is, um, you know, we, we've built MOD who we've
29:40
referred to in any number of ways, but she essentially is a harness.
29:44
And if we wanna put, um, you know, if we wanna put GPT 5.5 or we wanna put,
29:51
um, fable or we wanna put, um, an open weight model, you know, we can put
29:57
any of those models into the harness that we want and change what it costs
30:03
for the farmer to use the tool and, uh, what the tool is best suited for.
30:08
And certainly, as we all know, the orchestration of using the right model at
30:12
the right time is where the enterprise, um, enterprise rollout of AI is right now.
30:18
Um, it's clear that you're not going to use a model like Fable to
30:22
do, um, certain simple things with the weather and the soil, but, um,
30:28
there is a tremendous opportunity.
30:30
So MOD would hold in, uh, all of the, you know, information that we
30:34
think is critical and the way that we manage context is all in the harness.
30:39
Um, and then simply a reasoning layer.
30:42
Um, we bring that in, um, and, and sort of control where the data is
30:47
going out and what's coming in.
30:49
And so sometimes I'll query Mod and Slack and ask her a
30:51
question and she provide me much.
30:54
She can't provide me any strategic insight because she's on bare metal, she
31:00
has access to whatever data we fed her.
31:03
She could tell me what the conditions are on one of our test farms and what
31:08
that could mean for the farmer, but she certainly can't, um, you know,
31:12
help me with a grant application or, or help me, um, you know, help me come
31:17
up with a strategy for social media.
31:19
So, you know, that, that management of, of resources is, is where we're at.
31:24
And, and there's a lot that happens in there that, um, keeps,
31:28
you know, keeps our, our users insulated from the types of risks.
31:33
And, and that's a lot of potential as the sentinel layer.
31:37
You know, the sentinel layer is gonna end up being its own sort of a gentle
31:41
air, you know, it's basically the, the guys, the guys in the, in the guard
31:45
towers with the machine guns, you know, making sure that no data gets out that's
31:49
not supposed to get out and no data comes in that's supposed to come in.
31:52
So it's, I, I've all, I continue to advocate for.
31:56
You know, a national tiered architecture that will help us
31:59
solve a lot of these problems.
32:01
And I think that at the core of AI and agriculture in Canada right now,
32:06
it's an infrastructure question.
32:07
You know, the same as we need to have electricity and we think about 5G
32:11
and these resources going to farms.
32:14
I think that we need a collective national rollout of an architecture
32:19
that assesses the risks that Moss spoke of and uh, and mitigates them.
32:24
I think doing it collectively with a national role, provincial
32:28
roles and, and farmer roles, um, I think that's quite clear.
32:32
And, uh, I think that, um, the next policy framework and the federal
32:36
AI strategy need to support that with the serious investment of, uh,
32:41
hundreds of millions of dollars.
32:44
Jesse Hirsh: Well, and Jen, let me remind you at any point, just unmute
32:47
and I'm, I'm happy to let you in here.
32:50
I I Donald, to throw you a, a quick follow up past that and, and Mohsen,
32:55
I'd also like to hear your thoughts.
32:57
What about the cybersecurity side to this?
33:00
And, and I say that because you evoked the sentinel sort of protecting our data.
33:05
I actually, yesterday just used an LLM to hack into a device that I have that I
33:12
wanted to be able to access the data into.
33:14
So that was me in the positive sense, leveraging some of the
33:18
abilities for these LLMs to bypass any obstacles that get in their path.
33:24
But how concerned should we be about the cybersecurity of the systems we're
33:28
building, given that it feels that we are in, uh, an inflationary period when
33:34
it comes to cybersecurity capabilities?
33:37
Donald: I think extremely concerned.
33:39
I think it's another good reason to take a, a collective national
33:42
approach to building computational infrastructure around artificial
33:46
intelligence in agriculture.
33:49
And I think part of that, a architecture has to be, um, air gapped, um,
33:56
gapped, inference empowered, um, computing inside the farmer's field.
34:01
I, I don't think most of the data that farmers are collecting
34:05
doesn't need to leave the field.
34:07
I think it's, uh, it's, uh, we don't want the internet necessarily
34:12
connected to what farmers are doing.
34:15
Um, and so is where a lot of our research and development is right now.
34:20
Like, uh, we have sort of broad stroke ideas about how to implement this
34:25
nationally, but when we bring it back to the PEI Federation of Agriculture
34:29
and its members, um, you know, we run, we run a, you know, we run a hardware
34:35
stack that's designed and is being designed and experimented with to
34:39
ensure that, um, the, the, the core data from the farm is never in harm's way.
34:47
Um, and that's why the sentinel layer's so important.
34:50
It's the sentinel layer's not guarding on the outside of the, of the cloud.
34:55
The sentinel layer is actually guarding between the farm field and the cloud, so
34:59
that, um, so that the data is not, don't want things crossing that air gap that
35:06
you're not expecting, which is possible.
35:08
And, you know, hallucination is, is, I'm not sure hallucinations are the
35:12
issue they were in 2024, but, um, certainly data escaping, um, is something
35:20
that we, we don't wanna, we don't wanna, uh, we don't wanna play with.
35:24
So huge.
35:25
And, and there are people in Canada thinking very critically about that.
35:29
And we're lucky.
35:30
Um, cybersecurity on farm has been an issue with robot milkers and, and, you
35:35
know, all types of, I ot, you know, whole barns, and we see get delivered
35:40
and it's a much bigger issue than the, than the industry talks about.
35:45
It's one of those things that when it happens to a farmer, uh,
35:48
they prefer to keep it quiet.
35:49
So the scope of the issue currently is, um, broader than
35:53
people may, may understand.
35:56
And, uh, in addition to that, I mean, maybe the scariest thing about AI is
36:01
that they become better at cryptography than us and all of a sudden start
36:04
locking a set of our own systems, right?
36:06
And so we don't want to be like, that's the Terminator.
36:10
Like that's the danger, you know, like we, we, that's the danger is,
36:13
um, we wanna be a lot better at cryptography than these things.
36:17
And cryptography is the box that they live inside of.
36:21
Jesse Hirsh: Right on Mohsen.
36:23
Mohsen YN: Yeah, so I'm not expert in cybersecurity, but uh, my view, I think
36:29
like cybersecurity should be a very serious concern where we talk about
36:32
like, you know, AI and agriculture, it shouldn't like stop us from using ai.
36:39
But you know what?
36:40
In agriculture, I, I, I, I see cybersecurity is not only about
36:44
protecting computers or the data, is also protecting trust.
36:49
And the trust is the real foundation of the AI adoption
36:52
when it comes to the farmers.
36:54
Like this is, that is why I'm saying like, you know, cybersecurity is the very
36:59
serious concern in agriculture because if we are just breaking this kind of
37:04
trust, that's, that's not a good sign.
37:07
So as Donald said, some of the data, there is no need for the data to
37:11
go, goes outside of the barn like, you know, outside of the field.
37:14
So we can like, you know, develop so many different infrastructures or to
37:18
so many different, like, you know, ways and approaches handle this type of data.
37:22
But, and in like, you know, in the breeding, I can tell you that one way
37:27
that we are trying to, uh, increase the security of the data that we are
37:31
producing is to assign different numbers, like, you know, assign different IDs
37:35
and everything and protecting them from like, you know, being used by others.
37:39
And then we try to come up with a good, like, you know, who owns the data, where
37:43
we are going to store the data when we are, and how we are going to use the data.
37:47
We always have the version controls and we are securing all this
37:51
data, like, you know, in different infrastructure that university provided.
37:54
So are all the all different steps that we are taking to make sure that
38:00
the data that we are collecting and data that we are, that we have right
38:04
now are in a highest protection level.
38:07
But again, I see like, you know, cyber security as a very serious concern.
38:11
And yeah, we need to also build more infrastructures and more using more
38:16
approaches to make sure that the data that we're producing are secure and safe.
38:21
Jesse Hirsh: Well, and I appreciate you using the word trust because I think trust
38:25
is something that is often associated with, say, the adoption of AI and
38:31
certainly the federal government centred trust in their national AI strategy.
38:37
I feel both of you have already kind of addressed a lot of the prerequisites
38:42
that users kind of require for trust.
38:46
To what extent, to what extent is trust an ongoing moving target?
38:52
And and I say this because I kind of feel in today's conversation, we've
38:56
gone through leadership, we've gone through governance, we've gone through
39:00
kind of research and application.
39:03
But I think the bias once again that all of us have is
39:07
literacy and curiosity, right?
39:09
We like to learn new things, we like to explore these types of systems.
39:14
Not everyone has that.
39:16
A lot of people are just overwhelmed by the day to day and what they focus on.
39:21
I, I say this as a way to segue into the climate side of this discussion because
39:27
it strikes me that while we often talk about AI in terms of productivity, we
39:32
talk about AI in terms of research.
39:36
What are the areas that I think AI can be particularly powerful is managing
39:42
complexity and managing uncertainty.
39:45
And it, it strikes me that that's exactly what climate volatility and the
39:51
kind of climate crisis that I think the Super El Nino brings to us this year.
39:57
All of that is a grand transition that kind of asks the question, what
40:02
role do you think AI can play in responding to our changing climate?
40:08
And, and Mohsen, I'm gonna throw to you first 'cause I suspect you are well ahead
40:13
of the rest of us, uh, both in terms of anticipating the needs of the beans
40:19
that are gonna be growing in volatile environments, but also in terms of some
40:23
of the digital twinning and some of those simulations and research you do.
40:27
So, uh, taking this question from any angle you desire, where do
40:32
you see the role of AI in helping us respond to climate change?
40:36
Not just in a reactive sense, but in a proactive sense so that we
40:41
start actually feeling like this is a, a kind of crisis we're capable
40:45
of responding to effectively.
40:48
Mohsen YN: Yeah, no, that's, that's a very great question, Jess.
40:50
The point is in, I can start from the plant breeding side.
40:54
So, in plant breeding, we are trying to make a cross, and then we are advancing
40:59
the lines for the next eight to 10 years.
41:01
So we are selecting potential parental lines right now with the hope that the
41:06
children, the pro of these patterns are going to be superior in the next 10 years.
41:12
So you can't imagine, like you, we are seeing how like, you know, climate
41:17
has been changed over 10 years.
41:19
Right.
41:19
You know, we are seeing more frequent, like, you know, extreme
41:23
weather and that's, that's something that we was, we, we are not able
41:29
to try to see on the history.
41:31
We are not able to learn from the history.
41:34
It's something that we are facing something new every year, every year,
41:38
and weather is always ahead of us and surprising us this kind of climate change.
41:45
All this ai, I can tell you predictive ai, like all the predictive AI
41:51
models, are going to, are trained based on whatever data that we
41:55
are collecting from the field.
41:56
So we call them input and output.
41:59
So the output is actual yield, actual performance of plant, and the input
42:04
is the historical data, like the historical climate that we have, right?
42:08
So we are expecting to predict the performance of plant based on the
42:13
historical climate, environmental data that we collected different from different
42:18
stations, whether it's stations, we try to predict the performance these specific
42:24
plan for the future, where in the future we are not able to see what's going on,
42:32
what will happen in terms of the extreme weather, extreme, like events that happen.
42:39
one thing that we are doing, and we used to do, was to try to collect
42:44
some of the climate change scenarios.
42:47
All the data that is available based on different anthropogenic
42:51
climate change scenarios.
42:53
We, they call it RCP, representative concentration pathway.
42:57
They recently changed it to the SSH is socio society, something like that.
43:02
I, couldn't remember the, what is the actual acronym means, but it's SSH.
43:07
So there are different levels sorted from 2.4 down up to 8.5.
43:12
So 2.4 or five is the case that the climate change is very
43:17
mild and so good in the future.
43:20
So nothing is going to happen.
43:21
Like everything will be perfect in the future.
43:23
What would be the temperature in each location look like?
43:27
So scientists predicted that, and then climate like RCP 4.5 is
43:32
everything is exactly the same as now.
43:35
What will happen in terms of the climate, like, you know,
43:38
all the variables in the future.
43:40
And 8.5 is the worst case scenario.
43:43
If the fossil fuels will be bad, like, you know, we are burning more fossil
43:46
fuels and there are so many wars, there will be so many wars in the future.
43:50
So what would be the temperature in different location?
43:52
What would be the, uh, climate patterns in different location in the, like, you
43:56
the future up to 2000 and a hundred.
43:59
So we have this data.
44:01
We can also use this data as a testing in the ai, but again, pay attention that
44:07
we have output, which is the performance and the input, which is all these like
44:12
average daily data, whatever that we have.
44:16
So something that I've mentioned last time, and I wanted to also, if
44:21
there is a time I wanted to discuss about it, is the compute, like, you
44:24
know, uh, quantum, sorry, quantum ai.
44:28
in the quantum ai, the reason that I'm thinking that, uh, future belongs to the
44:35
quantum AI is because in quantum AI we are able to search more thoroughly, not
44:43
only looking at the input and output, but also try to find all these combinations.
44:49
Let me like, uh, provide a like clear example.
44:56
say that you are flipping a coin.
44:59
when you are flipping a coin, the normal computer or the, like the ai, predictive
45:04
ai, just counting the number of the times that the coin landed as a head or tail.
45:13
And it, at the end, when you just flip the coin, it'll tells you that in
45:19
what percentage, in what possibility you will see head or tail, right?
45:24
But how about only look at one time, one time that we are flipping coin.
45:31
Coin at each position has different like, you know, uh,
45:36
it's different possibilities.
45:39
One time is head, one time is coin.
45:40
Head coin, sorry.
45:42
Uh, head tails.
45:43
Head tails.
45:44
Right.
45:44
And then do you think that it's landed the possibilities 50
45:49
50, like 50% head, 50% tail.
45:53
about head tail?
45:55
Head tail and landed as head.
45:58
So we have four heads and three tails in one, one event that happened.
46:07
So quantum tried to capture kind of things in one, one data that we have.
46:15
no matter how many, like sometimes that we are seeing, AI is not able to predict
46:23
We are trying to collect more data.
46:26
how about using the best whatever that we have.
46:32
then in these kind of possibilities in agriculture, we have genes,
46:37
we have climate, we have, uh, aggro like agronomy practises.
46:40
We have so many areas and all these having so many combinations.
46:46
So the events that we are, the output that we are seeing in one
46:49
year is only one combination.
46:51
But how about considering all of them?
46:53
And this is something that we are trying to move toward that, like
46:58
capturing all these combinations.
47:00
Try to see how many of these combination has the highest possibility
47:04
that happen in the future based on the climate change scenarios.
47:08
And then the next step is to design digital twin.
47:12
So we are at the University of Guelph in different sectors in food science.
47:17
I know that, you know, I'm collaborating in some of the proposals.
47:20
They are using digital twins in plant breeding.
47:22
We already saw to use the digital twins.
47:25
And last year I saw to do that.
47:27
And I can tell you, uh, that all the crosses that I made, like
47:30
90% of the crosses that I made last year was because of the ai.
47:34
All the, it was the AI crosses and I'm waiting to see what would be,
47:38
I'm so like, can't wait to see what would be the result like in the field.
47:42
But anyway, we try to design digital twins.
47:47
Based on the facilities that we have, we have growth chambers and we can set
47:52
different temperatures set different, like, you know, conditions that will
47:55
happen, likely happen in the future.
47:57
And also investigate all this and assessing all these combinations of
48:03
the genes, environment, soil, all these kind of things in a very small scale.
48:08
And then test the highest possible one in the field and then make a selection and
48:13
select for the potential parental lines.
48:16
Maybe some of the panels that we are selecting today, their
48:20
performance is not good.
48:22
They are not providing a, like they're not good at all, but they
48:26
will be good in the future, in the next 10 years based on this climate.
48:30
So if we are using the conventional methods that we have, we will
48:34
definitely dis like, you know, discord, all these bad ones.
48:39
if we are adding computing like, you know, quantum, sorry, quantum
48:43
AI to that, we may able to find more like, you know, patterns.
48:49
We may able to like get the best juice again out of all the data that
48:54
we have and then we can make our selection more accurate and maybe
48:58
select something that is not good right now, but it'll good for tomorrow.
49:02
yeah, that was the strategy that we are taking here in terms of the research.
49:08
Jesse Hirsh: Well, and I am grateful again for you really helping me wrap
49:13
my head around the revolutionary potential of quantum computing,
49:16
especially when it comes to the wide ranging probabilities that any farmer
49:22
has to deal with when making decisions.
49:25
Uh, Donald, there was a lot there to build upon.
49:29
I, I do want to bring us back to the way in which, uh, climate and climate
49:33
volatility can be assisted by, by ai, but I suspect your mind was also blown, uh,
49:39
uh, by the quantum implications there.
49:41
So by all means, take it whatever direction you wish.
49:45
Mohsen YN: Uh, so sorry Donald, before you talk, I, I forgot to say, something that
49:50
like, you know, the credit of the quantum ai, I just wanted to give the credit to
49:55
people who bring that idea and we discuss.
49:58
So the credit of the quantum AI is, uh, Dr. Rosita Dora, she is computer
50:06
scientist at University of Guelph.
50:08
And we had, we, we, uh, we have AI for food in select, you know, initiatives
50:13
and we were all on the same table.
50:16
We had a lunch and she was talking about the quantum and we were
50:19
discussed this more throughly.
50:21
So I just wanted to also give the credit to Dr. Dora as well.
50:24
But yeah, please Donald, go ahead.
50:27
Donald: Yeah, I mean, the probability of the coin exists is almost a hundred
50:30
percent, you know, not quite, but, um, so it's quite a universe and, and when Jean
50:35
Jacque Russo wrote the social contract just before the French Revolution, he
50:39
probably wasn't aware that there's only, there's no less than a hundred percent
50:42
chance that the coin actually exists.
50:44
But nonetheless, you know, governance remains incredibly important.
50:48
Um, and, and the human element in it.
50:51
I mean, my thoughts on quantum are that, you know, it's kind of a tailwind to
50:56
this idea that the ram companies have quadrupled in value and that, um, and
51:00
that, you know, that the US economy should be investing $700 billion in, in compute.
51:06
You know, is there a path here where, where we can bring a tremendous amount
51:10
of compute to the, to the world, um, without necessarily having to
51:14
build these massive data centres.
51:16
Um, so yeah, I mean, we can see the venture capital space
51:19
around quantum is, is increasing.
51:21
I'm sure there's brilliant students at, um, at Waterloo who are at the absolute
51:27
forefront and, and I'm sure they're all very excited to move to the United
51:30
States to commercialise their technology.
51:34
so, uh, on the climate change piece, I mean.
51:38
Geez, I, I, I've been trying to make a carbon credit since I got to the
51:42
Federation, um, in 2021, and I couldn't move the ball forward on it until I had
51:49
the benefits of large language modelling.
51:51
And, uh, I mean, the, the protocols and the data requirements and the
51:56
documentation is, is simply too much for grassroots NGO, uh, without
52:01
the benefit of, uh, of all written history, uh, broken down and modelled.
52:07
And then when I feed it a VERA document, I can actually push the ball forward.
52:12
And so MOD found VT 0 0 0 1 4 for us, which, which allows us to, um, model soil
52:19
carbon the way that we wanna model it.
52:20
She also, um, found a number of different pathways.
52:25
I, I think, Jesse, when you talk about complexity and climate change,
52:28
these tools and their ability to, um, identify relationships between,
52:34
between data sets that we haven't previously considered is immense.
52:38
And I think that will, in a number of different ways bring value to farmers.
52:42
And, uh, the convergences between disparate data sets that these
52:46
tools unlock is very exciting and simply like the ability to create,
52:51
um, simulation environments.
52:53
You know, as soon as we fed mod the necessary data, we, we do now have
52:57
a, a 50 year bio geochemical model for Prince Edward Rhode Island.
53:01
That's sort of the first, uh, foundation from which the whole system runs.
53:05
And, uh, and our farmers are users benefited from it immediately.
53:09
So managing the climate and managing the impacts of climate change.
53:13
And, and, and additionally, I mean, we had a, a a once in a lifetime, or
53:17
what was a once in a lifetime drought last year in Prince Edward Island
53:20
that cost our economy $500 million, which is a, a big part of our economy.
53:24
And, and so I, I prompted Maud if she could sort of manage climate risk the
53:30
way that I've always wanted to manage climate risk and, and managing climate
53:35
risk and adapting to climate change has always been a messy, qualitative,
53:39
multi-stakeholder data, you know, qualitative data analysis problem.
53:44
And, uh, mod has sharpened it up dramatically identified the
53:49
key data sets, brought in the data that she has available.
53:52
And I believe very quickly we've, uh, built something that can ensure that
53:57
every dollar we invest in, in, um, irrigation in Prince Edward Island reduces
54:02
risk by, uh, by the maximum amount.
54:04
And, and so that's very exciting.
54:06
So on the mitigation side, I don't think you can make soil carbon credits
54:10
effectively without LLN technology.
54:13
And on the adaptation side, certainly managing complex hazard risk modelling
54:17
and, uh, and reducing that risk effectively with, uh, investment, I
54:22
think on adaptation and mitigation, um, as well as understanding when
54:26
the actual weather impacts come, how to, how to respond to those.
54:29
I think.
54:30
On all pieces.
54:32
You know, as someone who's very passionate about responding to climate
54:35
change, you know, this technology is certainly the biggest leap forward.
54:39
Um, you know, since the advent of the Post-It note.
55:15
Aggressively, um, you know, I mean the, the, you know, managing natural
55:23
resources is about engaging with all of the stakeholders effectively and
55:30
understanding how everyone values economic capital, social capital,
55:36
capital, and environmental capital.
55:38
Uh, it's up to the stakeholders to, um, identify the, how
55:44
they want to distribute that.
55:45
That's why governance capital is so important.
55:48
'cause it controls the distribution of those, um, capital buckets.
55:53
So we need good democracy to, democracy is and capitalism is ai, right?
56:00
Like, we need democracy to keep us safe and control our capitalism and, uh,
56:06
ensure that capital is being distributed between those three buckets according to
56:11
the, uh, the wishes of the stakeholders.
56:13
Stakeholders and, and, and, and waiting those stakeholders, you
56:17
know, is, is a challenge, but, uh, certainly not impossible.
56:23
And, and, and, uh, is a, is a wonderful sort of place to be, uh, to be working
56:28
to create a more sustainable world.
56:30
Mohsen YN: I can add to that.
56:31
So, you know, Jen, I can, I can't talk about this like topic for hours.
56:36
The point is need to be honest that AI has cost.
56:41
So there is no doubt on it, like AI has a cost, but the way that we are
56:44
using ai, it can just justify the cost.
56:48
Let's say I, as a plant reader, I make a selection for the future that,
56:54
and try to, let's say, using, I try, I try to use AI to release a call
57:00
to work that would be useful for the future in like, you know, in and
57:07
would be perfect and produce higher, higher yield in this climate change.
57:11
Right?
57:12
You know, all these volatile things that we are seeing right now.
57:15
So AI for developing this variety.
57:20
At the same time, let's imagine that I'm not using any AI and I don't have any
57:25
information that what will happen in the future, and I don't have the prediction
57:29
ability, predictive ability to do that.
57:32
So I'm releasing some varieties in the future that maybe those
57:35
varieties are not good and farmers need to add more fertilisers, more
57:39
pesticides, more herbicides to that.
57:43
And then so many of these kind of things are going to deplete to the
57:46
soil, spread to the environment in the future and the cost.
57:53
So 10 years, I, I'm right now using ai and of course, whatever I'm using,
58:00
has a cost of the water, using water carbons, all those kind of things.
58:05
And it's bad for the environment.
58:07
There is no doubt on it.
58:08
But let's imagine the war and the prediction that the, the, the
58:12
breathing without AI for the future.
58:15
How about consuming more fertilisers and those kind of things that's
58:20
even worse for the environment.
58:23
these two has the cost, but which one has the better cost is the use of AI in
58:28
this case also, when we are using ai, we need to come off with a good strategy.
58:35
That's something that I was just thinking for months and I couldn't find
58:40
a single answer to that is how we can better communicate with AI and encourage
58:46
people not to use AI in, for everything.
58:50
Like I'm seeing every emails, like more than 90% of the emails that I'm
58:55
receiving right now is written by AI or like, you know, formatted by ai.
59:00
Why?
59:01
What's the reason I wanted to make mistake?
59:04
I wanted to make mistake and writing some words, like having some typos
59:09
and everything is way better for the environment that we are living.
59:13
Even one emails is nothing for me, but let's compare it.
59:17
Let's consider it like for the, as a worldwide, how many emails
59:21
we are sending to AI to ref, like, rephrase, revise for us, right?
59:26
That's, that's something.
59:27
And even, you know, we are not paying attention to the actual
59:31
models that we are using.
59:32
Sometimes we are using like GPT five, oh 0.5, whatever the latest
59:37
one for revising one email.
59:40
So it's, I'm usually telling everyone that it's look like that you getting a very
59:45
expensive, fancy car and running in a, like a street with maximum 30 kilometres.
59:52
Why you wanted to do that?
59:55
That's, that's not even good.
59:56
And we need to have this kind of policy.
59:58
I was just thinking of developing a kind of software that can use us in between as
1:00:03
a middle, between like the user and ai.
1:00:07
And it can just like, you know, understand what kind of models your question
1:00:13
require and then use a specific model.
1:00:16
Or sometimes the way that we are talking to ai, no one trained us.
1:00:21
This is something that just came up we are all sort to adopt it
1:00:28
without fully understand how to better communicate with ai.
1:00:31
So I'm seeing that so many times we are writing a actual letter, Hey, I, hey, I,
1:00:37
I wanted this, this, this, this, this.
1:00:39
And then please do that and then add this part.
1:00:42
So is a very small example that I'm telling, but you can just see it on the
1:00:47
larger scale, how much we are wasting tokens, how much we are wasting energy.
1:00:52
So you can't believe that you know these type of emails that you are or this
1:00:58
type of text that you are sending to ai.
1:01:00
It has like AI need to like read first that cost energy
1:01:05
and then we call it intent.
1:01:07
Just keep all your intent that cost another energy.
1:01:10
And then reasoning of the ai, thinking of the AI that costs energy
1:01:15
and then provide you the response.
1:01:17
So how about keeping everything short?
1:01:19
Understand how to better communicate with AI in a very short and consuming very
1:01:25
less tokens, and then be efficient and then use AI wherever are really needed.
1:01:31
Not for everything, right?
1:01:33
That's something that we can do it.
1:01:35
But at the same time, in terms of the research, you should
1:01:38
see the cost and the balances.
1:01:40
Like with ai.
1:01:42
Without ai, what would be our future look like and what would
1:01:45
be the environmental cost?
1:01:48
Jesse Hirsh: And that's, I think, a brilliant frame,
1:01:51
uh, for us to both end on.
1:01:53
But think about our next, uh, panel, our next gathering, because as Mohsen
1:01:59
was kind of, I think, quite brilliantly sketching out the economics of, you
1:02:05
know, having a sports car, a super sports car to go out and get groceries
1:02:12
when you could walk, uh, to get the groceries, which would be far healthier.
1:02:16
Mohsen YN: We go.
1:02:17
Jesse Hirsh: Imagine that applied to greenhouses, because greenhouses are a
1:02:22
topic I'd like to talk about next time.
1:02:24
I think what I'll do is I will go out and get a guest from the greenhouse
1:02:28
industry, interview them as a solo episode on the podcast, and then
1:02:33
invite them back for our panel.
1:02:36
Because most, and what I was thinking about is, on the one hand, yes,
1:02:41
greenhouses are great for Canada.
1:02:43
They help us grow stuff, but they are resource intensive.
1:02:47
They are environmentally intensive.
1:02:50
Maybe there are some crops that don't belong in greenhouses that we
1:02:55
should be thinking about growing in less intensive growing environments.
1:03:00
Again, a question for a future episode, but I think if anything, what these
1:03:05
panels are allowing us to do is, on the one hand, make sure that Jen, your
1:03:11
literacy, your understanding of this is constantly upgraded because if anything,
1:03:16
you are, part of the reason we're doing this is to make sure that other leaders
1:03:21
within the sector can understand these questions and understand these issues.
1:03:26
But I don't know about you guys.
1:03:28
My questions tend to be increasing faster than my answers.
1:03:33
I leave every discussion we have with a whole bunch more topics that I want
1:03:38
to discuss, so I am grateful once again for your time and expertise and invite
1:03:45
you all back so we can keep having these conversations and really fostering the
1:03:49
kind of cross sectoral conversation that we're successfully having.
1:03:54
Mohsen YN: Thank you so much.
1:03:55
Yeah, I'm looking forward to that.
1:03:57
You know, every time that I'm coming here, I'm also learning at the same time.
1:03:59
Like from Donald, from Jen, from you, Jess.
1:04:02
I'm just learning and yeah, I would be happy to continue that discussion as well.
1:04:06
Yeah.
1:04:07
Jesse Hirsh: Right on.
1:04:08
Donald: Yeah,
1:04:08
Jesse Hirsh: you.
1:04:08
Donald: again soon, guys.
1:04:09
Have a great day.
1:04:11
Jesse Hirsh: Indeed, a great day was had by all certainly.
1:04:16
I hope you had a great day listening to that conversation.
1:04:20
Again, due to some of the technical issues, I've already listened
1:04:23
to that conversation many times.
1:04:26
Uh, something I normally loath to do when having to produce
1:04:30
an episode for the podcast.
1:04:32
But in this case, I quite enjoyed it.
1:04:34
There were multiple layers to the insights that Mohsen and Donald offered, and
1:04:40
again, my gratitude to them for taking time to engage here on the podcast.
1:04:46
And I think really elevate our understanding of ai, our, uh,
1:04:51
ability to see the vista, the imminent event, event horizon, uh,
1:04:57
i, I found incredibly valuable.
1:05:00
Now, as I've mentioned, AI has been a recurring thread,
1:05:05
uh, here on the Future Herd.
1:05:07
We've got an episode, uh, already recorded, not yet released, uh, with
1:05:11
Dr. Suresh Raja, a professor in AI and livestock at Dalhousie University.
1:05:19
That was a fantastic conversation that I'm eager to drop, and I'm working on a
1:05:24
virtual event, uh, for the future heard on open source and AI for mid-fall.
1:05:31
So keep an eye on that and I'll continue to drop hints every now and then.
1:05:36
Uh, thanks again, uh, for listening, for engaging for rating the, uh, podcast on
1:05:43
your favourite podcast network for sharing it with your friends or for sharing it
1:05:49
with your revolutionary cell plotting to destroy all data centres and enact the but
1:05:54
Leary and Gha D no matter where you lie on this particular, uh, zeitgeist issue.
1:06:01
We're glad that you've joined us and listened to what we've
1:06:04
had to say on the issue.
1:06:06
Uh, until next time, I am Jesse Hirsch and this is the Future Herd.