Zuzalu, Onwards
- 13 minsZuzalu, Onwards
Wow. Zuzalu happened and now I’m back home. Still digesting everything that happened there, while rejoining the same daily activities as before. But will it ever be the same? It’s incredible to consider how much can be catalyzed during a dense space-time event with people open to experiment. So much that it’s difficult to focus on the usual work/activity that must continue throughout such experiences. And that’s how I want to start this labnote. By saying that I’m late on this experiment’s schedule because of how things beautifully unfolded at Zuzalu. I’m working on a larger “Zuzalu reflections” piece of writing that explains what is behind this “delay”, but the more pragmatic aspects are:
According to the schedule, by now I should be running the focus experiment. That is, I should be connected at all times to a “peer-being” via the VHP. Well, that is not the case.
The VHP is not ready yet. Well, it is working and currently vibrating with temperature changes, but I haven’t managed to connect it directly to any bio sensor yet, mainly because of incompatible BLE protocols, as explained in the previous labnote. I’m currently investigating the possibility of it being connected to online biodata (like what’s coming from the e-plant), with the help of a mobile app. The advantage of that is being able to use the bracelet being far from the sensor - like when I go to the city - via data service from my phone carrier.
The good news is that I have found a “peer-being”. The peace lily (Spathiphyllum wallisii) has been sitting with us for a few months now and I have always been amazed by how sensitive it is to environmental changes, both in regards to light exposure and temperature. And it responds fast to these changes, allowing for important visual cues to be compared with the haptic signals I’ll receive in my body. (one very important element derived from sensory substitution experiments). I have also come to the conclusion that the source sensor for the focus experiment will be the one I’ve been using since the beginning of my tinkering with biodata. It is the sensor I understand the best, and that has given me more consistent results in comparison to the other ones I’m using. Attaching it to the Nicla Sense board seems like a good approach to continue monitoring environmental data (temp, hum, press, etc) and transmitting it via both BLE and WIFI. Besides, I suspect that the Nicla board can use the same BLE protocol as the nRF52’s one (the chip behind the VHP), which would make things much easier. I will also use two channels for the biodata reading. That is, two of the same sensors reading simultaneously in different plants of the plant. Below is a prep video for the setup, where I weave plant roots and electronic wires:
[[root/wire setup video]]
New time horizon
So when will the focus experiment begin? My new current target date is the next full moon. June 4th, running for a full moon until July the 3rd. How about the next itens in the timeline? Well, I’ll start by stating that this project is definitely extending the 4-month period I had originally imagined for it. Not only because of three unplanned weeks dedicated to Zuzalu, but also because I totally fell for the planning fallacy when estimating how much time it’d take to develop some of the deliverables. That is, the writing pieces (in the same of a paper and a short-story) and an end-user friendly SymBioWare version. I’m more than ever committed to their delivery, but I also want to make sure they will be in the best shape possible, which will certainly require more time, especially for the beta device.
In practice, all of these things will be developed during the focus experiment - I will be journaling thought it, which will provide a lot of material for writing pieces. And I’m sure I’ll be constantly improving the tech, especially for the data analysis, which will be paramount to any market version. I’m just not feeling comfortable to establish a new data for when these will be ready. I prefer committing to continue posting regularly about their updates.
Beside all of this, one thing is clear: this work is growing like mycelium under-ground and new aspects of it are emerging all of the time, in a way that makes it a real challenge to fit it into any regular project management framework. Just to give an example of what I mean, while at Zuzalu I helped birth one of such hyphae: Sentient, giving voice to the Morača River (soon to be declared the first river-based national park of Montenegro). With the demo we built in a “nature hackathon”, humans are able to chat with the Morača ai chatbot, accessing information from the scientific papers used to feed it. We presented it at the Zuzalu town-hall, for the entire community, and such was the positive reception that we’ve decided to continue working on it (given our individual capacities).
But more than a cool project, Sentient opened up our imagination to what kinds of realities would be possible if a LLM model (like ChatGPT) could be designed to speak in the name of a non-human entity, like a river. This could be done with the help of local communities who inhabit and care for the river, with all of their stories and symbolism around it, and grounded on live sensory data that could help determine the entity’s health and wellbeing. This can be a powerful combination, especially when considering the Rights of Nature movement and natural entities like rivers, and even individual animals or bee hives being granted legal personhood in different parts of the world. Literally, it felt like we’re in the beginning of something grander.
From here, onwards.
Zuzalu
Curiously, the whole experience brought me back memories of the period I lived at Minerva university. Perhaps it was the multi nationalities represented. And/or the fact that almost everyone there had a digital nomadic life, spending way more time in front of a computer that what I’m used to. The wealth? People’s age? Or perhaps the sensation that in both experiences it took a considerable amount of effort to “blend in” with the local landscape and communities - for language and behavioral barriers. Regardless, the Zuzalu experience felt… familiar. With the difference though that now I had a family of my own to go back to, which really made things different. Despite all of the exciting things happening there, I wanted to go see my daughter the moment I landed.
Besides the “saudade”, I opened up. I made myself present, engaged with the community through varied actives - from football, to hacking - offered my contributions via capoeira sessions, clean-ups and active listening.
At all moments I asked myself how the interspecies work I’m developing could interface with the people and activities there. The initial idea of issuing Zuzalu passports to non-human citizens didn’t seem to attract much attention and the “synthetic biology week” was the closest thing to the “growing of symbiotic senses” in the official schedule. But I arrived right after it. Nonetheless, one could smell in the air an unofficial emerging theme around “other than human intelligence”. Such a perception was, of course, backed by the more-than official “Artificial Intelligence” track, but also reinforced by other smaller spontaneous trends. For example, I arrived in the middle of the “Zero Knowledge” hackathon and was surprised by the concept of “autonomous worlds”, being explored by at least two teams. I loosely joined one of them, also because it was the team led by the person who brought me to Zuzalu (Thank you, Danilo!) They were developing an on-chain game whose world changed autonomously based on a ML model. BabyGaia what the name of the game. My main contribution was to add realtime biodata (from the e-plant sitting at my lab) to one of the characters behavior. That’d provide an interesting platform to experiment with machine and organic inputs in a game setting.
The two pieces were made. The game MVP was functional and I managed to ChatGPT a python script that would pull the real time biodata upon request. But we didn’t manage to get them working together in time for the hackathon presentation. And curiously, the will/interest to put integrate the two pieces faded away after this event - mostly because of the new incoming inputs from the larger Zuzalu experience, I’d say.
But the simple mention of such a “possible future” integration to the larger community was enough to somewhat trigger the collective imagination around “other than human intelligence”. During the very same presentation I received a message from another zuzalan, Noah, wanting to further explore the intersection between the biodata coming from my lab and artificial intelligence.
In the few days right after the presentation I still believed that I could do both Zuzalu and “normal work” (in my case, the symbiotic senses research). It was necessary a complete shut-down of all equipments I had brought for me to realize that that was not gonna work. After dealing with the frustration of not being able to set up sensors to collect biodata locally, I began exploring other routes, like feeding the live biodata coming from my lab to ChatGPT and prompting it to role-play a plant experiencing the conditions described by the data. It was a fun and illustrative exercise, although not very accurate, as ChatGPT can’t really make sense of the data. It would say things like “my roots are a little bit more electric than usual” but with zero parameters to compare the upcoming data against.
But this silly exercise gained a whole lot of context when two zuzalans, Jasmine and Marijana, invited the community to brainstorm around how emerging technologies could help amplify environmental protection.
In that session a novel universe of exploration was born. With a lot of enthusiasm from the attending crowd, the ai-chatbot-to-provide-voice-to-natural-entities idea sparked the imagination of many of us. I was one of the most enthusiastic one, as I had literally just done something very similar to that in my computer.
A “nature hackathon” emerged, with people wanting to “hack” on the river’s data. Is there available data? Curiously, a very special event was scheduled in the zuzalu program for that very week: a rafting trip to the Tara river. One of the most pristine - and one of the last wild rivers - of the world (also referred to as the “Tear of Europe”), it called us!
So there we went, in an unforgettable trip to one of the coldest waters that I have experiences - thus, coldest night ever. The emerald of the river’s water is just as breathtaking as uncapturable. Rafting on it was such a thrilling experience! Feeling the river’s movement as it flows. All of the power and surprise, accompanied by awe and wonder! The second largest canyon of the world! Wow. At the same time, sad feelings for seeing that it has also been trashed by the human kind. There are no places left from our oil fingerprint it seems.
And after rafting, a very mysterious experience of hearing the river whispering on my ears, inviting me to go inside it despite my warmth and comfortable situation. I couldn’t refuse. Cold plunging in it was such a difficult to describe experience. I mostly remember the feeling of burning skin, and cold bones. But it wasn’t uncomfortable or painful. In fact, it felt like cleansing. I silly toe bumping into one of the rocs opened a small hole in my toe. A lot of blood running out. No pain. Foot rivering.
Back from mystery and wonder, the 3-day-long nature hackathon took some shape with two teams with shared membranes. I was focused on continuing the work I had done with the e-plant, but this time getting the algorithm to role-play a specific river, powered by a smaller subset of information about the river. Based on what we could gather from science partners - scientific papers and little raw live data - the approach was to train the model on the text from these research. This was amazing experience for me to come closer to current “AI” technologies. From both painstakingly working with chatGPT’s API and embeds to being exposed to state of the art solutions that take that work away. Shout out to retune.so for having made a really annoying process so easy approaching. What I worked on for 3 days with partial success (the model would not role-play the river, despite being prompt to), was achieved in 30 minutes via their platform. Totally worth the investment. Not to mention DJ’s openness to hear us and help us out.
After all of this work, a project was born: sentient, giving voice to the Morača River (for its position as soon to be declared the first river-based national park of Montenegro). In our demo, humans are able to chat with the Morača ai chatbot, accessing information from the scientific papers used to feed it. We presented it at the Zuzalu town-hall, for the entire community. Big shout out to Gordon, for the incredible video created around the mythology of the river and its life carrying nature.
Then again the “other than human intelligence” vibe was increased and the demo received a lot of positive feedback! From that moment on it was clear to the people who worked on it during the nature hackathon that they would want to see it developing further somehow. Based on each one of us’ capacities, the Sentient project gained tow co-founders (Jasmine and Marijana), two advisors (Noah, and Remi) and a tech-weaver (me). It is currently taking its baby-steps as a legal entity, with the initial value proposition of providing AI chatbots for non-human entities.
Sentient opened up our imagination to what kinds of realities would be possible if a LLM model (like ChatGPT) could be designed to speak in the name of a non-human entity, like a river. This could be done with the help of local communities who inhabit and care for the river, with all of their stories and symbolism around it, and grounded on live sensory data that could help determine the entity’s health and wellbeing. This can be a powerful combination, especially when considering the Rights of Nature movement and natural entities like rivers, and even individual animals or bee hives being granted legal personhood in different parts of the world. Literally, it felt like we’re in the beginning of something grander.
And that’s how I’ve been feeling lately, with renewed imagination for what’s possible with the history behind me. Zuzalu has opened perceptions and connections to new allies on this journey. I hope to be able to keep up honoring these opportunities.