Does Open Source AgTech make sense?
October 12th, 2022 - 22nd Edition of Easy Observations in Ag
Visit my website www.easyobservations.com for more info about me or follow me on LinkedIn and Twitter
Opening Thoughts:
This one sort of came to me last minute though I think the timing is relevant due to some recent news going around of a new AgTech open source initiative that I’ll talk about it a bit here. Before I dive into that I figured I can break down a little on my general open source thoughts.
For one, I have zero issue with open source and neither should you. It is literally the backbone in some ways for basically every tech based solution out there. Some more than others naturally, but it is critical. My dad built Satshot completely off of open source code and so have many other AgTech systems beyond their own custom development.
This is not to say that it’s always the best thing to do though. Traditionally as I’ve seen any type of open source code out there usually needs a good deal of customization to work specifically for the Ag industry. Some have done this better than others I can tell you that. The issue is Ag is just weird in how it works especially when it comes to mapping, the types of data files, and the editing/analyzing that data requires.
What is changing though are some new and old players coming to the table that are starting to create more specific open source based solutions for AgTech specifically. That is good news surely, but I do see some potential issues as well and I’m going to dig into that here.
A Synopsis:
I’m sure I’ll miss a few other protocols out there in the AgTech Open Source world, but I wanted to showcase a few I’m aware of that seem to be very interesting. I’m not going to dive too deep into them each, but yeah a synopsis.
While each of these likely have some similarities they also seem to be quite different. Also, they might be using some of each others stuff for all I know. It wouldn’t surprise me. It is always common for synergies to happen in Open Source, but it can get weird too as competition is competition. I do see that AgStack does have some similar connections below with OATS Purdue and FarmOS, but not sure to what extent. Partnerships can always be good, but they don’t always work either. Time will tell I guess.
The beast in the room, Microsoft. As many probably know they are also working on their Farmbeats platform. I’ll be honest here, I haven’t seen much directly on what Farmbeats does beyond the typical “we want to connect everything” montra and I don’t know many really using it. Heck, Microsoft’s second biggest office is in Fargo in my backdoor so I’ve heard plenty on it though I haven’t ever heard much of groups using it. What I have heard is that it is…not what many in North America are looking for. There are other stories, but I won’t go there. In any realm, clearly the Farmvibes.AI and Farmbeats stuff is connected. They recently just opened it up to and it’s why I thought I writing this so there is that. It does make me question somewhat of the larger play in what they are doing, but hard to say.
Purdue has been part of the Open Source stuff for awhile actually. If you remember, especially if you were a part of AgGateway like I was, Purdue was behind the semi similar solution called OADA. They did get some partners, but to be honest I have zero idea if anyone is using what they have or not. I know not everyone is using AgGateway’s ADAPT framework either so there is that too. Regardless, they seem to have some cool things they are building and they also seem very practical to many. Purdue as an educational institution is a giant in the Ag world as many know so they also have some power behind them. They have some connection to AgStack below too so that is interesting as well.
Linux is the beast of Open Source code as many know so them getting into the AgTech game is very interesting. They have a huge community, though I’m not sure how many people in Ag are in it. Like I said, Ag is different and you have to have people that understand it to develop things for it being technical or anything else. They have some huge partners with IBM and Hewlett Packard and other people (including some of the ones mentioned above and below) working on it so clearly they have some power in pushing this narrative. It does beg the question to me though, do they have too many partners working on it? Being involved in AgGateway for some time I can tell you giant partners mixed with many others can create issues. Stuff can get done, but the politics at times can get messy at times.
FarmOS is an interesting smaller group for sure. I don’t know much about them, but they seem to be getting something done. There is the AgStack connection and also seems some Purdue collaboration as well so they are getting some traction. As with any open source process the community is such a needed part and it looks like they have some decent traction there as well. They’ve been around a decent while I feel too so that proves there is something there of value. I’m sure they’ll continue it whatever success they are getting.
OpenATK is a new one and while it doesn’t look like they are doing everything they are going after some of the simpler protocols. The Tillage and Rock app they have are sort of cool as they are just needed items that everyone needs. That is what I see Open Source doing well in the AgTech world. Taking away the burden of rebuilding every wheel. If these guys stay going after the simpler apps I think they have a good future ahead of them. Worth a try for sure.
Do we really need it?
Yes and no and also I’m not sure. What I do know is that the idea has even crossed my mind in regard to my families company Satshot. There are parts of it I know many could use as others are just spending so much time building the same thing over and over and it doesn’t get the attention they want. It’s not to say I wanted to let it all go, but just certain parts as I know they work and are stable. Those parts are nothing really special, they are good standard processes in working with imagery in Ag I’ll say. That was the thought at least.
Part of me still wants to do that, but I also don’t because the thing about Open Source is it needs a good community and partners to work. Especially this is true in Ag. It is not to say I couldn’t figure this out in some respect, but it is a lot of work. It takes a weird determination to that not all have. This is why you see Universities in the mix in AgTech Open Source and tech giants. They have the connection, time, money, and power to push through it. Plus a plethora of grads that give free work for the universities. Also Microsoft and Linux are who they are…Yeah.
Either way, I mention this as it showcases some of the issues with Open Source AgTech. One wants to do it to help others in the industry, but you also don’t want to give up everything so as not to create a bigger competitor. It depends on the one doing it too. This is where the Microsoft thing sort of scares me. I get it and it’s great and all, but there is also much more to what they are trying to do. Heck, Tesla let go many of their patents to be open and while on the surface it looks like a great thing to do, it likely was also done so people use your processes in development but fail to bring it to mass production so then Tesla can charge others that are too deep in R&D with consulting fees to make that tech or open source knowledge to actually work.
There is always something missing in Open Source that needs some added intervention I’ve found. Nothing surprising about that as it can’t be perfect for everything everywhere, but sometimes learning just the open source code that is available takes a special person to work with it anyway. Let’s talk about GDAL as an example. A staple of the geospatial open source world. You use it everyday likely. The thing is, not everyone knows how to work well with GDAL. You need good people that work with it and understand geospatial coding processes. That is the thing with Open Source. You still need the expert that understands it or may be part of that community.
Now you have your next issue, community. These AgTech Open Source groups sound great, but what is the community like. Clearly the Microsoft one is literally Microsoft people at this point and yes many program in their world, not everyone does either so as to traditionally avoid them and their license fees. Linux can almost be this way too and really is popular due to Microsoft and their license fees. Linux is a fee version of Microsoft OS more or less, but you still need people that understand it to manipulate and work with it. Yes, Linux has a super huge community but how will that work with Ag types?
Here is the biggest issue I see though, many in AgTech that have built software that would use these open tools already have what they need built. Newer similar groups are not spouting up like they used to either. Heck, there are only so many players that would even need it anyway regardless and then they’d have to add to that community too giving their generally small competitor market similar tools they have themselves. Why give someone else your unique needs? The AgTech space is just not as big for some of these Open Source protocols yet as some may think.
In Closing:
While I said I’m not against any of these Open protocols in AgTech, I think they have some limits and things they will have to deal with. As I just said, I’m not sure their are enough out there at this point in the market that really need these tools anymore. They’ve already built some of them out of necessity as there was not Open Source for them before. You can’t just wait for it.
Also as I’ve said many times, AgTech groups need to be niche and specific to really work. These Open Source tools are broad as they should. Niche doesn’t always need broad, though sometimes they do. Basically, it is not going to be for everyone but will be for some. That is the problem too. If you need a community to help build the Open Source protocol than you need a decent group. If you only can grab a small group then it will be more problematic to keep the community engaged in the shared process. This becomes even harder if the Open Source code being showcased is something multiple competitors use.
If AgTech was like other Tech industries then I’d say Open Source would have a much better path overall. Is that possible in the future, to an extent for sure. There are definitely needs for Open Source in the world of tracking, shipping, food waste and safety, and on and on that do not seem to really exist yet that many if not all need. That is more of the successful path I see in the need for AgTech Open Source. Not the Precision Ag semi-Agronomic space that many seem to have built so far, but more of the Agrifood space that deals with safety and traceability of sorts. That’s a huge community.
In all, I hope all these groups and any others I didn’t mention can make an impact and dent in the AgTech or just general Agriculture world. I really do. If I can help in someway I’m in as well. There are many of us out there and ones reading this that should get more involved in Open Source AgTech projects. I mean, many of your companies couldn’t exist without it anyway so many it is time to give back too. I’ll leave that to you, but overall more is needed. So, to answer my question though in do we need it the answer is YES. It needs more time I’ll say and a bigger community while making sure they don’t alienate the users of the open stuff itself.
It’s tricky more or less. Time will tell…
Thanks for reading Easy Observations! Please Share, Subscribe, and Comment if you would be so kind. I’d also be happy to get together and meet if you’d like to talk more in-depth about AgTech or Earth Observation in Ag. Feel free to contact me at my website. All the best!
Does Open Source AgTech make sense?
We use Apache and Linux on all server architecture. This allows smaller firms to run on "Bare-Metal servers" and retain full control of their service. There really is no other choice unless you allow yourself to be handcuffed by AWS, Azure, and the like. Ironically there is a higher cost for doing things "the right way" using open-source in this situation. Cloud services offer a much lower barrier to entry, with scalability which appears built in, but if you read the fine print, costs absolutely skyrocket with scale.
We use ODKCollect to do our cocoa purchasing/export and quality system in Belize. We also use it to track vehicle usage, growth in the field of our vanilla and heaps of other stuff. Open source, works great where there is bad connectivity (Central America, Africa), good community of support and easy to program. Agree there needs to be standards so my quality language = yours.