🌑 Pathfinders Newmoonsletter, November 2024

This new moon, we wonder how AI agents will deal with cookie consent banners, explore both human and AI conversational powers – and responsibilities –, and invite you to visit a library.

As the moon completes another orbit around Earth, the Pathfinders Newmoonsletter rises in your inbox to inspire collective pathfinding towards better tech futures.

We sync our monthly reflections to the lunar cycle as a reminder of our place in the Universe and a commonality we share across timezones and places we inhabit. New moon nights are dark and hence the perfect time to gaze into the stars and set new intentions.

With this Newmoonsletter, crafted around the Tethix campfire, we invite you to join other Pathfinders as we reflect on celestial movements in tech in the previous lunar cycle, water our ETHOS Gardens, and plant seeds of intentions for the new cycle that begins today.

Tethix Weather Report

🕶️ Current conditions: Despite clouds getting heavier, you can still get sunburns when looking at all the bright-mirrored reality distortion fields emanating from Silicon Valley, so keep some low-tech sunglasses within reach

Fire practitioners are feeling so optimistic about our collective future, they are taking a break from prepping their doomsday bunkers to spread the gospel of AI salvation with us, mere mortals and web crawlers. Well, unless the web crawler belongs to their competitor, in which case they have to protect their business interests and brilliant ideas from being scraped and added to the AI commons. (Hint: Try asking OpenAI’s ChatGPT to read the essay Machines of Loving Grace written by the CEO of Anthropic.)

But we’re all in this together, the techno-optimists with escape plans reassure us. AI will “fix climate” any day now, so don’t let the biodiversity warnings and news of failing carbon sinks get you down. The fire practitioners promise to offset our collective carbon sins and build nuclear power plants to power our cities data centers, so we can all keep feeling good about our (over)reliance on the increasingly heavy clouds. (See: Humanity is on the verge of ‘shattering Earth’s natural limits’, say experts in biodiversity warning, Trees and land absorbed almost no CO2 last year. Is nature’s carbon sink failing?, Google Backs New Nuclear Plants to Power AI, and Amazon signs agreements for innovative nuclear energy projects to address growing energy demands)

Surely, you will rejoice at the news that the brightest AI golem minds are now being trained to relieve you of the drudgery of filling out boring forms and spreadsheets. And why, yes, they might also relive you of your job, but another good news is that the fire practitioners at OpenAI have now appointed a chief economist to study the impacts of your redundancy. (See: Claude AI tool can now carry out jobs such as filling forms and booking trips, says creator, Google is reportedly developing a ‘computer-using agent’ AI system, and Dr. Ronnie Chatterji named OpenAI’s first Chief Economist)

But again, you’ve got nothing to worry about. When you try to find a new job, an AI golem hired to manage human resources will be right there for you to tell you just how inadequate you are at filling forms. Inevitably, you’ll pay your own AI golem to fill job-application forms for you, so that the AI golem on the other end will evaluate your golem’s outputs. Rinse and repeat, until it’s AI golems processing forms all the way down, and all humans fall out of the loop. (See: Job hunting is demoralising enough without having my personality eviscerated by an AI interviewer and Welcome to the Ministry of Futility)

Before we reach that point, we hope we’ll realize AI golems actually make decent brainstorming and conversational learning partners, as we recently explored in our podcast. Perhaps they can even help us remember what all these forms really are for. And reconsider whether form-filling really is worth the energy and resources that could otherwise power a small city.

If you’re wondering and pondering about these questions, we hope you’ll join us in the climate-focused generative AI Exploratorium throughout November. The Exploratorium will be a collaborative online exploration journey in which we’re giving ourselves the time and space to uncover and process diverse perspectives through open conversations.

And if you’re looking for additional seeds to inspire both explorations and conversations, we hope that the seeds we have collected for this Newmoonsletter help you look for paths not yet illuminated by profit-seeking fire practitioners.

Tethix Elemental seeds

Fire seeds to stoke your Practice

It’s been almost two years since ChatGPT took the world by storm. And now, AI models are getting ready to step out of the familiar boundaries of chat boxes and API requests, and venture into the wild world web. Get ready to start seeing more LinkedIn posts about agentic AI as our AI assistants learn how to use the mouse and figure out how to get past cookie notices and captchas.

And if you are struggling to keep up with all the AI buzzwords and hype, look no further than the State of AI Report 2024. The report is a lengthy recap of this year’s developments and trends in AI research, industry, politics, safety, and more.

Just as interesting as keeping track of the evolving AI capabilities is examining the shifts in AI-related myths and metaphors that shape our mental models of how the technology works. This blog post explores how the AI metaphors of toolbox, companion, and enchanted object distribute agency differently between the user and the AI.

As the metaphor of AI as an agent – hopefully not of chaos – starts gaining traction, it is certain to bring its own mental baggage to the evolving mythology of AI. And a recent NN/g blog post rightly brings attention to the fact that emerging technologies offer fertile ground for tech myth-making, which in turn affects how people interact with new technologies. Especially ones that aren’t fully transparent about their inner workings.

And while it appears that most of the tech industry is focused on adding new AI capabilities and releasing new devices, real innovation is done by engineers who are figuring out how to do more with less. Take for instance /e/OS, a “deGoogled” version of Android that also extends the lifespan of devices no longer supported by their manufacturers. With a leaner OS, even older mobile devices can happily run for longer, helping their owners reduce their environmental footprint.

We would all get more out of our existing devices if the software we use – and build – wasn’t so bloated. This includes our websites. If you work in web or are curious about learning more about the building blocks of the web, we invite you to explore HTML for People. This beginner-friendly online book can remind you that you don’t need a complicated tech stack with multiple dependencies to build and publish a website. Just plain good ol’ HTML for content and structure, and a bit of CSS for style.

And hey, these days, AI assistants can also help you simplify your projects and ideas, you just have to ask them the right questions. For inspiration, check out these examples of how you can quickly build simple web tools using Claude’s Artifacts feature, mostly with basic HTML, CSS, and just a sprinkle of JavaScript.

Air seeds to improve the flow of Collaboration

Perhaps pairing with AI coding assistants might help us build more efficient software. Maybe even help us figure out how to “green” other types of roles. Or are we fooling ourselves and generative AI is just a resource-heavy distraction that’s moving us further away from each other and our climate goals? Between Sam Altman’s naive insistence that AI will “fix climate” and failed calls for an AI pause, lies a sea of nuance and diverse perspectives that don’t get as much coverage.

Throughout November, Alja will be facilitating a Generative AI Exploratorium within the ClimateAction.tech (CAT) community that invites you to take a deep dive into the sea of nuances with other curious explorers. The format of the Exploratorium is based on our previous work on collaborative online learning experiences, but adapted to accommodate different modes and levels of participation.

If you’re interested in joining this climate-focused Exploratorium, we also encourage you to join the CAT community for full access to all resources and the Exploratorium Slack channel in a wonderful online community. And if the times of the four live discussion sessions of the Exploratorium don’t get along with your timezone or work schedule, you can still participate fully async or even organize a local discussion group. Reliable sources tell us that Mat is itching to start an APAC yarn circle.

And if you can’t find any human discussion partners, you might even enlist the conversational skills of your neighborly AI chatbot. While the reliability and omniscience of AI chatbots are still very much up for discussion (rightly so), AI chatbots can make good conversational and learning partners.

Perhaps they might even help us retrain our innate conversational learning abilities, as we explored in the most recent episode of our podcast. Especially if we’re willing to let go of the answer-machine metaphor and focus more on being a conversational partner. (Which includes remembering that sometimes we all say things we overheard somewhere without bothering to fact-check.)

Conversations help us learn, make sense of the world, and even inform our decisions. Unfortunately, there is a dark side to every great power, as the latest suicide-related chatbot lawsuit reminds us. With great conversational power comes great possibility, but also great responsibility. Unfortunately, the latter is a lesson tech companies aren’t yet taking that seriously in their mad rush towards profitability and market dominance.

Earth seeds to ground you in Research

Recent research shows how collaborative imagination of a shared future can bring people closer together in the present. We suspect a similar effect might be observed in human-AI co-imagination, despite the AI partner just playing along with whatever the human needs. But the resulting human bond might be just as strong, which is why chatbot relationships can have such a strong hold over humans, occasionally even leading to harm.

That’s why the way we design and integrate AI capabilities – especially conversational ones – into products matters. For instance, an AI mediation tool built by a Google DeepMind researchers showed how iteratively generating statements that best reflect the diverse viewpoints of a group can help the group find common ground, even on divisive topics. This approach works because humans retain the agency of reviewing and rating statements.

When considering adding Large Language Models (LLMs) to our products, we must remain mindful and vigilant about the limitations and quirks of this technology. From security concerns to hallucinations in medical transcripts, LLMs still appear to have a “mind” of their own.

Recent fairness research from OpenAI also shows they are not completely immune to first-person biases based on seemingly trivial details like your name. And just like some humans, they also appear to be easily distracted by other unimportant details, as demonstrated by a team of researchers at Apple.

Perhaps different approaches that would enable agents to think fast and slow with a Talker-Reasoner architecture might better equip our AI assistants for dealing with the complexity of human demands.

And in more applicable research news, Google DeepMind and Hugging Face have released SynthID Text, a tool for watermarking and detecting LLM-generated text. At this point, it’s worth reminding ourselves that OpenAI decided against releasing their own watermarking a tool over fears of losing users, but also valid concerns about stigmatization of non-native speakers. As with most things related to generative AI, caution is always advised during implementation and deployment.

Water seeds to deepen your Reflection

These days, it can be hard to remember that digital technologies we now take for granted aren’t even that old. Take for instance search. Google would still be considered a young adult at its 26 years, but even those of us who are one or several decades older now struggle to imagine doing our knowledge jobs without a search engine. But this reflection from a former reference librarian is a good reminder that libraries and librarians met our (re)search needs quite well before Google. Obviously, we can now get answers to many more questions in much less time, but often by sacrificing accuracy.

Are we now headed in a similar direction with AI assistants? Time will tell. Well, at least until we are collectively willing and able to dedicate enough energy and other resources to keep the cloud and the rest of our computing infrastructure running. Eventually, physical limits will likely force us to rediscover and reinvent low tech solutions.

And to further demonstrate that technological “solutions” don’t have to be complicated, here’s some recent research which indicates that playing natural sounds to ecosystems could help them bring back to life by attracting different species. As with any technology, even this simple acoustic enrichment trick should be wielded responsibly to avoid attracting species to non-viable ecosystems or fooling ourselves into thinking that we – or an AI model trained on our biases – know best what an ecosystem needs to heal.

A music seed to sing & dance along

And now, for some simpler human acoustic enrichment, here’s a music seed that can remind us of the hauntingly beautiful fragility of our humanity. Modernity often expects us to be “good machines”, but at the end of the day, we are only, but wonderfully, human. Let’s treat each other with the care and kindness we deserve.

I can turn it on
Be a good machine
I can hold the weight of worlds if that's what you need
Be your everything

I can do it
I can do it
I'll get through it

But I'm only human
And I bleed when I fall down
I'm only human
And I crash and I break down
Your words in my head, knives in my heart
You build me up and then I fall apart
'Cause I'm only human

Pathfinders Podcast

If you’d like to keep exploring the lunacy of tech with us, we invite you to listen and subscribe to the Pathfinders Podcast wherever you get your podcasts. The podcast is a meandering exploration inspired by the seeds planted in the Newmoonsletter at the beginning of the lunation cycle, and the paths illuminated during the Full Moon Gathering.

The question that emerged in the October Newmoonsletter and guided our discussion was: Why aren’t more people using AI conversational interfaces for conversational learning? In this episode, we wonder whether digital technologies can satisfy our yearning for a yarn and support embodied conversational learning, especially through voice-based conversational interfaces.

You can watch the full episode on YouTube, or listen to it on Substack or in your favorite podcast app.

Take a listen and join us at the next Full Moon Gathering if you’d like to illuminate additional paths for our next episode!

Your turn, Pathfinders.

Join us for the Pathfinders Full Moon Gathering

In this lunation cycle, we’re inviting Pathfinders to gather around our virtual campfire to explore the question: How intelligent do AI companions actually need to be? – but it’s quite likely that our discussion will take other meandering turns as well.

So, pack your curiosity, moral imagination, and smiles, and join us around the virtual campfire for our next 🌕 Pathfinders Full Moon Gathering on Friday, November 15 at 7PM AEDT / 9AM CET, when the moon will once again be illuminated by the sun.

This is a free and casual open discussion, but please be sure to sign up so that we can lug an appropriate number of logs around the virtual campfire. And yes, friends who don’t have the attention span for the Newmoonsletter are also welcome, as long as they reserve their seat on the logs.

Keep on finding paths on your own

If you can’t make it to our Full Moon Pathfinding session, we still invite you to make your own! If anything emerges while reading this Newmoonsletter, write it down. You can keep these reflections for yourself or share them with others. If it feels right, find the Reply button – or comment on this post – and share your reflections with us. We’d love to feature Pathfinders reflections in upcoming Newmoonsletters and explore even more diverse perspectives.

And if you’ve enjoyed this Newmoonsletter or perhaps even cracked a smile, we’d appreciate it if you shared it with your friends and colleagues.

The next Newmoonsletter will rise again during the next new moon. Until then, find some time and space for conversations, visit your local library, and be mindful about the seeds of intention you plant and the stories you tell. There’s magic in both.

With 🙂 from the Tethix campfire,
Alja and Mat

---