210 – Two Years of The IDEMS Podcast: The IDEMS Collaboratories and CommonTech

[00:00:07] David: Hi and welcome to the IDEMS Podcast. I’m David Stern, a founding director of IDEMS, and it’s my pleasure to be here today with Kate Fleming, another director. Kate, it’s our two year anniversary episode. This is a big deal.

[00:00:20] Kate: Hi, David. Yes, it is. I also would like to call attention to your new microphone. I think you’re sounding great. Very high quality audio. 

[00:00:30] David: Well, it is your new microphone, which you’ve given to me to help the podcast in its effort to become more professional.

[00:00:38] Kate: It’s working.

[00:00:40] David: Great. And it is, it’s been quite a journey getting here to having a professional podcast microphone, to having more invited speakers and many more actually lined up. So this should be something which becomes more routine. There’s a real sense, for me anyway, that this is maybe soon going to be for an external audience rather than intended primarily as an internal audience, which is quite exciting.

[00:01:05] Kate: And there is an archive for people to get through, but it’s all worth it.

[00:01:12] David: Absolutely. And you know, going through all 210 of them at this point is maybe not everybody’s idea of fun, but we can use certain ones to highlight bits of work that we are doing. We can now draw on these, this archive as you put it, to highlight specific pieces of work, specific bits of thinking, it’s a real resource.

[00:01:34] Kate: Yes, and that is the plan, and I think we’ve talked about this before, we have more malleable use of podcasts, it’s not like we want people to sit down and listen to 200 podcasts. It’s much more you’re interested in a specific topic and now you can dive in, and actually, there is a wide range of ways that you could enter. Whether through, you know, if you’re interested in a particular aspect of technology, if you’re interested in an area of impact, if you’re interested in our team or people we work with, I think there will be lots of ground for us to cover.

And that is why as we were talking about what are we going to talk about on this podcast, it does feel actually like there are many different ways we could go where I’m not entirely sure. 

[00:02:17] David: And that’s what’s sort of exciting about this episode, but also in general where we are right now. It is a really, it feels fresh and new, it’s still hard work and there’s lots going on and it’s difficult to make ends meet and all these challenges. We’re actually at a place where, well, when I say we, it’s really you, you are actually at a place where you can tell our story, I tried this out recently and it worked!

[00:02:44] Kate: Well, you’ve done better than I have. I still struggle with our elevator pitch. I get asked, this has come up on our team too, what do you do? And I always do that. Uh, okay, well, and so yes, we are getting better and better. I think the positive thing was that we got to the place internally where we had kind of the long form, yes, these are how all the pieces fit together.

So now we’re in that phase of how do we chop this down? How do we do this sufficiently? How do we take out the pieces that people don’t really need to know? And that’s actually easier than when you’re looking at a story and being like, God, this still has such gaps. These just don’t connect and I don’t quite understand what’s missing.

And so I think that has felt very satisfying, we see the picture, the big picture. We see how all the pieces fit in. And yeah, hopefully, I don’t know, maybe in the next, maybe in 10 episodes, we’ll finally be to the point where we can try out our elevator pitch on our many podcast listeners.

[00:03:43] David: Well, let me get down to the key breakthrough, because what’s interesting is we’ve had coherence since day one, but we weren’t able to express this and bring anyone on the journey with us because it was just too complicated. And really the insight, there’s many insights you’ve had, but the insight to frame this as the IDEMS Collaboratory and CommonTech as two distinct components, which are both needed and intertwined in what we do, and both hold complexity, but really we are the interlink between those two. This really works.

And I can say it works because in a recent meeting where I introduced it to someone, after introducing in a relatively short space of time, they said, oh, that makes sense. It took me so much by surprise, I almost fell off my chair. It actually makes sense.

[00:04:39] Kate: It’s funny because I still think, one, I would say no matter what technology space you’re working in, this is such a common problem of you often have technologists who are thinking about, these aren’t even technologist issues, you could have social scientists, you could have anyone where you sort of understand a system, you’re thinking about the complexity, I think in many ways I sat from a very different perspective thinking about these same issues.

But the work of really distilling things down, of figuring out what is information that people need, what is information they don’t need, how do we kind of conceptualise these things in a way that people can wrap their arms and heads around? That is really, it’s hard. And I think also things can feel, you know, there were things we explored that just felt really esoteric, we were like, oh, this is great, and that is just like, there’s so much I would need to understand to even begin to meet you where you are.

So then it’s like, okay, back to the drawing board. And I would still say, I would even say the CommonTech, we’re not actually telling that much about what it is, but I would say every time we do a little bit more work on something and we think something’s landed, and then we take it a bit further, we have a new insight where it’s like, ah, that element of that wasn’t quite right, but this piece is right.

I suppose what takes so long is that it is a lot of you just have to commit to something, do the work, see it through, and only then can you realise if it works or doesn’t work, and that it has to be fixed. And that can feel like a long journey.

[00:06:13] David: But what you are describing and what I find so interesting is you are just describing the problem of communicating it. Because what’s so interesting is we’ve been going now almost eight years, and the vision that we had at the beginning is the vision that we are actually enacting now, and that you came from a totally different path, but with the same vision.

What’s so interesting is that, actually, the core of what we’re doing hasn’t changed, it hasn’t moved, we’ve deepened our ability to communicate it. But what’s so surprising to me is how few people have been able to latch on because we haven’t been able to tell the story right.

And that’s what I think is changing. And it might also be, you know, it’s of its time, that the current AI boom or bubble, depending on how you want to frame it, is focusing people’s minds on the fact that, oh, technology maybe isn’t the solution on its own, we need humans in the loop. We don’t want to lose, we could go on a sort of particular education path where we don’t want to lose teachers, teachers are really valuable.

But if you go too far down the education loop just with the AI as a technology, then you are taking teachers out of that loop, and you are trying to say you can get better education just with AI, which is not what we or many others want to frame. And so if you are wanting to make that case that we want the technology to be enhancing, lifting up those other roles, the other human interactions, not just the teacher interactions, but peer to peer student interactions, maybe student interactions across context, all these other things that could be of value, that needs a different type of tech.

[00:08:01] Kate: I think that is a lot of what we have bumped up against, that the conventions of tech, our product is technology, you have failed at developing a technology if you have built in humans needing to do things. It’s like the tech should just be set it, forget it, you go on your computer, you just use it, whatever that stuff is. And I think the fact that we always have had humans in the loop, and I use quotes around that as kind of an expression that we’re increasingly using, that it’s important for humans to be in the loop. That that is not a flaw, that is a feature.

And I think that is hard for people to wrap their heads around because it feels like an inefficiency, and it feels like we’ve done something wrong, where in fact this is what makes it work, this is what makes it accountable. And I think there are conversations happening about, particularly with AI, the goal should be human machine collaboration, not humans are replaced by machines. And also that’s not really the world we want anyway.

So what is this sort of human machine collaboration? And I think that was kind of a breakthrough insight for us because when you put humans in the loop and you don’t have AI, it feels more problematic for people, whereas alongside AI, somehow people start to be like, oh yeah, that makes sense that you would want that collaboration.

[00:09:26] David: And what’s really interesting is, exactly as you’ve expressed it, the inclusion of the AI component to help people understand the importance is really essential in what we’re now discussing and how we’re able to communicate this. But the value of humans and socio-technological innovation, this has been well documented in Africa. We can’t take claim for inventing this, it’s just that we have been exposed to another way of thinking about technology.

The classic example, which keeps coming back time and time again, is mobile money, creating so many jobs, creating opportunities, making buses safer. And it’s all about that interaction between the very human element, creating human connections, with the technology. And this has been known, documented, understood in these very low resource environments. But because there’s so much money being accumulated by the big tech approach, which tries to remove humans from the loop, that sort of distorted the narrative on this.

And my favourite example, which pulled the curtain back on this was Amazon Fresh, walk out of the supermarket, AI will just say what you’ve got, you don’t need to interact with another human. Until it was found out there were whole centers in India where people were watching you shop and trying to figure out what you had picked up, what you had left in your basket, what you had taken out of your basket. And it was actually their job to do that. And now they were paying someone not in your local community, but someone very far away, and not visible.

[00:11:14] Kate: But I would argue that is the classic tech play, which is fake it till you make it. And there are examples of people being accused of fraud for this. I would say something like Theranos, that’s held up as this emblematic fraud. But I’m like, I don’t see how that’s different from a lot of tech promise, just because it involved health and the stakes were higher. But I think that ethos is the Silicon Valley ethos.

So Amazon Fresh doing that wasn’t because they didn’t plan to eventually use computer vision to manage everything with AI, it’s just they weren’t there yet. Also, it’s part of a broader vision to price things differently. Who pays more for that can of chickpeas than this other person who pays a little bit less, what will the market bear? Which is deeply, I think so dystopian and more and more people are starting to grasp like, oh, this is what happens when we just seed everything to technology.

Anyway, just back to that Amazon Fresh example, it feels like there’s more in that.

[00:12:17] David: And you are right, it isn’t that there was not an intention to remove humans from the loop. That was always the narrative, and that was maybe the goal. What is interesting is, if I understand correctly, it was recognised internally that that was so far away that the algorithms were so bad and were really unable to do that, that that was deprioritised. And so it was accepted that for the foreseeable future, this was going to be a human task.

And that was not made public. This is, I think what you’re saying and this is what I think you are right, these are the instances where it is bordering on fraud because what your marketing as being one thing is actually something else.

[00:13:03] Kate: And it’s also selling a future that is a long way away and using that distant future as an excuse to absolve yourself of responsibility to workers in your community. This idea, basically what I hear in that is you are outsourcing even the most menial of jobs, like a job that should be accessible to somebody, which is to just be a, checkout person in a store. How do we just continue to get rid of these jobs, consolidate wealth in the hands of few?

And I mean, all of this stuff is, every time we have a conversation, unfortunately, everything comes back to money and incentives. What is this world that money is investing in and what is it creating and is it serving everybody or is it just serving a few? And we struggle with this, what money seeks out is often not what we’re doing, where the funding for us is not always obvious or easily accessed.

[00:14:08] David: And let’s be clear on what you’re saying here, because, as a small business, we’ve done fine on contracts. But as a tech startup, we’ve totally failed to actually present ourselves or get recognised as something worth investing in.

This is why the story is important, and what is so powerful about the expression of us as this combination between the Collaboratory and the CommonTech, that we recognise, and we’ve recognised since day one, we need to be both the service provider in some sense, and the ambitious tech startup. And marry those two with a very clear social vision and plan.

This is what we recognised from day one. But it’s only now that I think we are able to start telling the story of the technology and how the problem that it’s trying to solve has been identified not by a single Collaboratory, but by a whole set of different Collaboratories where we’ve identified, ah, this is the common set of problems where the current technology stacks are actually really failing.

And then what we’ve done is we’ve articulated, ah, there’s a reason that the current technologies cannot serve the sort of impact oriented focus that we want because actually there’s a hard math problem behind this. Of course, you know how happy I was when I found that out. 

[00:15:45] Kate: But I realised that even from my perspective, the people who were always the most helpful were the deeply mathematical minds, where they’re like, oh yeah, that would be a hard problem. 

[00:15:55] David: Yeah, this actually isn’t solved by what people are currently doing, yeah, however clever the AI is, it’s not looking at it in the right way, the data’s in the wrong format, we need to rethink this, we need to think about what is, anyway, I shouldn’t get lost on that. But it is, there’s a hard math problem behind this.

[00:16:13] Kate: Yes, and I would say that each Collaboratory is focused on a category. So it’s education, mathematics education, or agroecology, or public health. And so within each category, so within each collaboration there is software, there are things that already exist. So there’s an element of how do we curate things that already exist for this particular problem set.

And that’s unique to a collaboration problem. It’s the underpinning issues around data and interoperability and information silos and modelling silos. It’s these problems that, you know, what happens if I need my public health collaboration to actually be able to interact with my mathematics collaboration because I realised there’s a link there.

And right now that would be virtually impossible, it’s not virtually impossible to do, it would be literally impossible to do in any efficient, quick, you would have to do some massive funding exercise, get all these people on board, and even then…

[00:17:17] David: And it wouldn’t be worth it because the collaboration, it’s opportunistic, there is an opportunity for these to interplay. I just go very specifically into that case, the really interesting thing I’ve got fascinated by there, in the public health side, we have these parenting programs which reduce violence against children, have all these other positive benefits. And one of the things that they do is they try to engage with parents of teenagers.

And one of the areas that parents really care about in many low resource environments for parents of teenagers is they really care about education. And the biggest problem their kids are having is their maths education. And so here is an obvious entry point where both of these could be really enhanced if they could be brought together.

This is in the crowdfunding that SAMI is doing, Supporting African Maths Initiative, that big goal is to actually integrate the parenting programs with the education programs because, actually, both are enhanced, this is a win-win. But it’s not trivial to do. As you say, this requires work at a really low level to be able to build structures which enable that interoperability to happen across these silos.

But that’s the vision which has to come together. If we have really impactful social initiatives, we need to think, well, how do they work together. The group at Oxford, what they made their name for is these multi-arm randomised control trials where you actually don’t just look at individual interventions, but interventions which are put together in ways that you might not expect and to see does that have further impact by having them together? It’s really sensible, but it’s not easy technologically.

[00:19:03] Kate: It’s not easy technologically, and also you don’t want to have to depend on Oxford or the Gates Foundation or some massive body, you know, the slow moving wheels of that machine and its funding to make that happen. Where you could, if that were all quite easy, and it would probably happen in many places, these little pilot trials would come up in local places where you’d have someone who has that insight. They do small prototyping, small pilots, they’re just running it, they can just do it with a local university. You could end up with really valuable evidence of impact, I mean, with just a few thousand dollars even, you could imagine that something like that could happen.

[00:19:45] David: If the infrastructure was there, we have concrete examples of this, we have this farmer federation in Niger, one of the poorest countries in the world, who have built their own app, they have their own technology for their farmers, serving their farmers. If they could now slot an education initiative or supporting parents, a caregiving initiative, into their existing technology, they would grab it because it serves their community. It’s what they want to do.

But we are not there yet in being able to offer them that technology, being able to offer that level of interoperability. But it’s the sort of thing which we have partners on the ground, that’s just one of many, but they are ready for this and we’re not, we need to build out the technologies which they could then combine and integrate with.

And this is where what I think we are able to articulate now in a way that wasn’t possible before is that to do this, to create this new deeply interoperable technologies, we need the cutting edge research in artificial intelligence. Actually, maybe I shouldn’t say cutting edge, it’s certainly not the bleeding edge because the bleeding edge at the moment is going off to, you know, generalised artificial intelligence, that’s totally irrelevant.

What we need is the stuff which has just come, which already exist, it’s these large language models, which can now be put into small language models, which can be made accessible without needing huge elements of resources. It is very recent advances, but it is things which are available with the technologies that exist now as of a few years. And that is enough.

[00:21:29] Kate: And I think what becomes hard, even as I’m listening to you talk, and this is what we have struggled with, is so much of our thinking is best illustrated in these systems. But we have to reduce everything down to products. I mean, the amount of money that even just a simple intervention can get, we know that the ambitions, that the implications of the things we’re thinking about, and even the things that we have prototyped, that we have piloted, they’re out there, they’re running, they just need an injection of investment to build them out, to make them easily picked up by others, all of those kinds of things.

But when you are too grand, when it is too systemic, it is too hard for people to grasp and it feels like unicorns and rainbows and this is fiction. And so it has been a lot of how do we distill, or not even distill, how do we break some of the pieces out into discreet, smaller elements that are constrained, that are contained, where people can say, okay, yes, I understand this, and we’ve been doing this exercise recently for math education, STEM education I would say, where a piece of it is STACK’s, open source, online assessment, technology, and that in many ways is foundational. And if I had to put something at the center, I’d say STACK is a big piece.

But then we need to talk about AI authoring agents and the fact that we need other open source technologies to start to make STACK’s core functionality packageable into textbooks. And then we have to talk about version control of different variations, all these different things. And those are things we talk about everywhere, but through this single lens, I think we’re seeing, okay, this becomes more manageable for people.

We’re not getting them to think about how math education relates to a parenting intervention. We’re just getting them to think about, even within math education, what would it mean if teachers, low skill technologists, they don’t need to have coding skills or anything, what would it look like if they could work with an AI agent to develop their own course materials, to develop their textbook, to give every student a personalised lesson that’s relevant to their context?

Those are examples where you start to see, oh, this is how what you’re envisioning plays out. And then I think once people get that case, that application, then they can start to think like, oh, and if this were connectable to something else, that would be cool. But we’re not trying to sell them, you know, the world at first pass. 

[00:24:17] David: The key point there, and this is something where as a small organisation, small social enterprise, promising these huge, ambitious ideas just seems totally unrealistic. And it is, unless you take this Collaboratory approach. The whole point is we are not intending to hold everything. We, as a social enterprise in this mix, are part of the Collaboratories, we serve the Collaboratories. Our role is to sometimes deliver product, it’s sometimes to deliver service, but it is always in collaboration, Collaborative by Nature.

This is also something very difficult to express how being collaborative can outcompete being competitive. There’s some good evolutionary biology about this, so it does actually play out that way many times. But it is something which is a hard sell, if you want, to investors who are used to you being selfishly interested in your own piece of the pie as opposed to thinking about yourself as a collaborator, enabling a system to work.

[00:25:27] Kate: Yeah. And I think that relates to IP issues, and as you’re saying that proprietary thinking. I think I did a slide for some presentation we did on our Oxford collaboration, and I showed how Jamie at Oxford is pretty much equivalent to a CEO if you were just running a traditional structure, and Laurie, who works for Parenting for Lifelong Health, she’s kind of that Chief Revenue Officer, the person who’s out drumming up new business and selling people on it.

All of these people who are just the best at what they do are coming together, because it’s a collaboration, you literally are getting this insanely talented person who’s in charge of their role, that if you were just a startup and you tried to hire this talent, one, you could never hire it because they wouldn’t be interested in working for you, and two, I don’t know, the work of just trying to put together this team that starts to generate leads and get users and all this stuff, it’s like kindergarten compared to you’ve got this, like, PhD holding superstar team that’s just excelling in their field.

And so all these people are able to do what they do best, and when you’re putting that into a collaboration where everyone’s working toward impact, and so you are more collaborative by nature, you don’t just hoard your little piece of the pie in your little lane, it just unlocks so much more potential.

And I will say, even as I’m saying that, what we see is that these things that we’re doing in low resource contexts, there are sustainable business models that are emerging, we see that these things have value, because they’re unsolved problems in high resource context too, it’s just that there’s such a bunch of noise that it’s harder to see sometimes that they’re unsolved.

[00:27:11] David: Very specifically, just a week or two ago, I was in a meeting with Jamie, who you just mentioned, who was saying how after a presentation of what we were doing about the apps and chatbots in these low resource environments, he had all these high resource contexts coming to him and saying, can we do that in the US? Can we do that in the UK? Can we build from this?

And there of course there are business models which come in, which enable you to implement slightly differently, it’s always intended to be blended, and so maybe they can afford, instead of $6 per person to go for something which is 50 or a hundred dollars per person, where then there is a different interaction, there’s a different business model around this.

But it’s something where those opportunities then build out in ways which can sustain the whole, and which everybody is on board with that sustaining the whole is about having the right set of actors who can seize those opportunities and feed back into the system, enabling everyone to grow as part of this.

[00:28:14] Kate: And it is funny because I think we, well, I think both you and I stay the course, we both really believe in this, not just believe in it in some, you know, quixotic way, but really believe in it because we can see, we’ve worked in the spaces, I know that this would be transformative, I know that there’s demand.

But it is so hard often. This is when it comes back to these are communication problems in some way, you think about the reductive ways that people package things that capture the sensibility, that make people jump. I mean, artificial intelligence is a great example of this where it’s like you package this, you just have sold this narrative, you go out and you pitch it, and then there’s all this excitement around it.

And I think so much of what we are trying to carve out in this space is how do we package this in a way that is accessible and exciting, and we don’t want it to be fictional hype, but a lot of AI isn’t fictional hype, there’s a lot of good stuff there that really has substance to work with. And so how do we find that very succinct articulation that just makes people say, ‘ooh’.

[00:29:25] David: More than just getting people to say, ooh, what we recognize we need is, although we need to do the individual components, we also need people to get excited by, oh, this is more than just individual components. A good funder for us doesn’t say, oh, we like what you’re doing on the education piece, you should drop everything else and just focus on that, or we like what you’re doing with the parenting, you should just focus on that.

A really good partner for us says, ah, we like what you’re doing on the parenting, but we see that it’s contributing to something bigger and we are interested in that as well. And that’s hard.

[00:30:06] Kate: And I think that has been a challenge that we bumped up against, that we have one piece that’s an enabling piece, it’s interesting, but its function is derived from how it works in conjunction with other things. So we need people to understand the system, the layers, the interplay of those different functionality pieces, some of which are very wonky and some of which are quite not wonky, they’re very simple. 

[00:30:31] David: It’s just deep mathematics, deep mathematics is deeply serious. You can’t just dismiss it as being wonky.

[00:30:38] Kate: Yes, yes. But yeah, I think that has been a challenge. And actually it’s interesting because one of our partners is the Topos Institute, who we work with, and they are research and deeply mathematical, and there is actually a clear lane for their work in a different way because they just are, and I don’t say just in a minimizing way, their work is great, I love them, they’re really lovely and smart and great. But, you know, they’re in this lane where the funding channels exist and they’re having money kind of, I wouldn’t say thrown at them, but there’s interest in what they’re doing. 

[00:31:14] David: Because it’s academic excellence rather than actually this bridge between what they’re doing, which is really good theoretical work, which we understand and we see the value of, and making it useful. That bridge into the usefulness, if you’re trying to make it useful, why are you worrying about that theoretical stuff? No, we need both, we need that theory. That’s what’s enabling us to do safeguarded AI properly, to actually build things which are going to be safe. If you don’t have the deep theory, it’s not gonna work.

[00:31:50] Kate: And I would say Topos is quite aware that they need the applied examples, it’s the reason I got invited to a conference about, which I think is hilarious, that I got invited to a Caltech Mathematics conference on collaborative modelling because I’m actually somebody who has been out thinking about, well, how does this actually work in practice? How do people do this? How do people navigate what are often social dynamics, to bring things into coherence?

If you have the technology, but people won’t use it because you haven’t overcome political constraints or social conventions or whatever those things are, you know, great, it works in theory, but it doesn’t work in practice. And so Topos is very aware that they have to constantly be moving between applied and theoretical. It’s just that they don’t necessarily have to tell that story. They can, when it serves them, they can just stick in their theoretical, deeply mathematical lane and you know, the people who love that want to fund it.

So, yeah, I forget why I went down that side-channel, but…

[00:32:55] David: I think the reason you went down that route is that our story is particularly hard to tell because we tie into their word, the Topos world and the deeply mathematical approach that they’re taking, and yet we are rooted in our Collaboratories, which are impact focused. And people don’t tend to bridge these words. Our role is that bridge, able to see, create the sort of environments, the Collaboratories, the technologies that are needed to support those Collaboratories without understanding that that deep mathematics is needed to change the way technology is developed.

And that story has been one that we’ve been struggling to tell. But it is one which I think is becoming increasingly visible as the stories around AI, and the value of recent large language models is being recognised, but also it is clear that the money and the way money is being pumped into that right now is not right.

These trillions of dollars going into these data centers is the wrong way to pursue it. You need to have more efficient models, not bigger data centers. They’re already consuming too many environmental resources, this is not what AI should be looking at.

[00:34:16] Kate: And I think we’re also up against a narrative that, well, a bit of what seems like an oxymoron is that we’re getting more technical to become more accessible. That is something that has felt very, I think, incompatible to people. Well, how could you possibly be going into deeper mathematics to make something that’s more accessible to like a farmer in rural Niger. 

[00:34:41] David: I’m afraid as a mathematician, it’s obvious. Just because something is more mathematically complicated doesn’t mean it doesn’t make things simpler. I could spend a whole nother podcast explaining how algebraic geometry works and how integers are really hard, but if you get to projective complex space, then, actually, things are easy again.

So being more deeply mathematical does not mean things are getting harder. It can mean that things actually come together. That very explicit example, two lines meet, how often does sort of two lines or two curves meet? Well, if you are in complex projective space, it’s easy. Two lines meet exactly once and if you have a polynomial, then it’s the degree of the polynomial, which determines how often they meet. Try and do that with integers and you have no idea.

What do lines correspond to? That’s actually a harder definition. When do they meet? Well, this is when the equations get solved together, but actually integer solutions are notoriously hard to find. This is deep number theory. You are now into a really hard problem on the same question, and yet the more complex mathematics actually makes it easier. This may not be common knowledge, but it’s so obvious.

[00:35:57] Kate: Well, and I guess I would also draw the distinction between, you can engineer this supercar that’s really easy for someone to use, but when it breaks down, oh, it’s gonna cost you $50,000 to get the right part, and you’re totally dependent on us. So I would say another piece that is counterintuitive for us is not only that you could build something that’s more complex and have it be more efficient for users, but also that it would be easier for them to fix, easier for them to build, whatever those things are, that is really the step I think we’ve taken that is the most counterintuitive.

[00:36:32] David: Oh no, you are right. That particular step, that’s at the heart of what we’re trying to do. This is exactly why current technology doesn’t work, because you have the skills barrier to fix it, to adapt it, to own it. And that’s what we need to take, that’s what we need the deep maths for. The deep maths is what’s enabling us to put the tools to fix it, to own it in the hands of people who don’t have deep tech skills.

[00:36:59] Kate: Yeah.

[00:37:00] David: This is a good place to finish.

The one thing I did want to just mention is that, you know, we’ve reached the two year mark and one of the things which is exciting about what’s happening next for the podcast series is that we are going to start releasing in French as well. And so to finish on a word about that and sort of saying, watch this space, especially if you’re francophone and in the West African region and interested in agroecology, because that would be the initial focus of the francophone IDEMS podcast.

[00:37:33] Kate: Nice. And I would also add that maybe Michele and Chiara will do an Italian podcast. That has nothing to do with our particular work other than the fact that we have two Italians on the team who might enjoy that.

[00:37:45] David: The Italian episode or the Italian series might have to wait a bit, we’re pretty much at capacity trying to launch a whole nother series in French, but that will be happening by the new year in preparation and celebration of the 20th anniversary of the West African Community of Practice for the Global Collaboration for Resilient Food Systems. Yeah, it’s gonna be celebrating that and it’ll start next year, and hopefully it’ll become an established part of the IDEMS podcast series.

[00:38:17] Kate: Great, I look forward to it. I’ll practice my French.

[00:38:20] David: Thanks.

[00:38:22] Kate: Thanks David.