More-than-human aesthetics ⊗ Enchanted knowledge objects in LLM UI ⊗ Native Americans guarded against tyranny

Thanks to everyone who wrote in last week, not only will it surely help with the sp4m issues, but it led to some lovely discussions.

Just for fun, since this is issue No.333 and a large part of my archive has still not been transferred to Ghost Pro after the move, I thought I’d fish out issues No.3 and No.33 for a bit of a look back in time.

More-than-human aesthetics

James Bridle interviewed for a Japanese art magazine on the topic of more than human aesthetics. I haven’t read their Ways of Being yet but it seems to be quite aligned with where my own interest is drawn; as an invitation to explore other intelligences, to better understand our own and that of other beings. It’s also an opportunity to decentre ourselves. Bridle also talks about machine intelligence and how it can accompany humans, collaborate with humans. The way they frame LLMs within the systems they emerge from and how all our discussions and ways of seeing intelligences are boxed in by capitalism, empire, and language is well worth the read. There are a lot of great passages in there, see the highlights below, but I’d like to draw out this one:

The chairs that I’ve built by asking the AI to help me build chairs are the sort of chairs that are imagined by someone who has only ever read about chairs. They’ve never sat in one, which is a very important thing to understand about how AI understands things.

I just think it’s a great expression of what and how machine intelligence “understands.” I’m keeping it for future discussions.

When we talk about intelligence, we mostly, subconsciously or consciously, mean what humans do. So all of our thinking about AI is contained within this very narrow idea that AI is somehow mimicking some or all aspects of what we consider to be human intelligence. […]

I’ve always found it incredibly striking that AI, this big amorphous conceptual thing, is having such a huge moment in the popular imagination just as in a lot of other ways, in the art world and elsewhere in politics, in our relationship to the planet, we’re also rethinking the centrality of the human. […]

We are living within the consequences of that imperial, colonial, epistemological action. Whatever anyone else in the world thinks, we’ve set the world up to run as though there’s nothing but the human and really nothing but certain types of human that matter. So the project, the only important project at the present, is to change that situation. […]

So even there, immediately, this other aesthetic, a non-human aesthetic, is bigger, broader, deeper, wider, stranger than a human aesthetic because it extends into all these different potential realizations, awarenesses, and therefore, things to process and therefore, things to think with. It increases the number of objects to think with.

Enchanted knowledge objects in LLM UI

It’s been almost four years since a great post by Jay Springett, Discord, DAOs, and the DWeb, and it still comes to mind once in a while, I still wish Discord had built it. This new piece on knowledge objects in the prompting of AI and how they could be part of the UI goes right along with it and will also be popping in my head with regularity, I expect.

In this framing, “knowledge objects” are “a book, or big PDF [dropped] into an LLM’s context window, it has a sort of gravity — like how objects with high mass reshape space-time — drawing the model toward specific ideas and reshaping the entire context landscape.” If you’ve done quite a bit of prompting and conversing with ChatGPT and the like to write texts and/or extract information from them, you’ll recognise this effect, these kinds of “objects.”

Jay goes on in more detail in his piece, on how they could be seen as talismans and used in various ways for more advanced prompting, have a look and I’ll tease out three things. First is his mention of using highlights in a text to signpost important information, i.e. draw the attention of the AI to what you are interested in. That’s something I’ve been trying/hoping to do with custom prompts for summaries in Reader, I think it has quite a bit of potential. Second, and I think it’s brilliant, Springett added markings within a long document, using ꙮ (the many eye glyph) for that same drawing of attention, including instructions to the LLM at the top of the text. Love it! Third is a prompt (pun intended) to other experimenters reading, I’ve looked a bit into Fabric and I’m pretty sure the kind of things Jay is talking about would be “easily” done within this framework.

Anyone who’s used NotebookLM or held lengthy conversations with an LLM about their own writing has probably noticed that whilst grasping the full context of a document, they can sometimes fixate on throwaway lines or inconsequential details. They often zero in on parts of a text that aren’t central to the argument or message at all. […]

We can Enchant Knowledge Objects (books, websites, PDFs, whatever) with power—guiding the AI’s focus within documents through structured annotations and symbols and turn them into powerful Talismans. […]

I see Knowledge Objects as handcrafted artefacts—collections of dense metadata and symbols that play with the model’s context in unpredictable ways. Weird, markdown files full of material that, that when dropped into a model, produce something entirely different than expected: shifting tone, reordering context, or amplifying particular ideas.

How Native Americans guarded their societies against tyranny

À propos of nothing (ah!), this piece drew my attention. In her article, Kathleen DuVal highlights how Native American societies developed intricate systems of checks and balances to guard against tyranny, similar to those later adopted by the U.S. founders. Contrary to the (sadly) common perception of Native Americans as primitive, many communities, such as the Iroquois Confederacy, practiced consensus democracy that prioritised shared decision-making and egalitarian values. During times of crisis, like the onset of the Little Ice Age, these societies decentralised power, creating more democratic structures to avoid oppressive leadership. DuVal emphasises that the ideals of leadership in these communities aimed for calm deliberation and the well-being of future generations, offering valuable lessons for contemporary governance.

As I was reading, I realised that I shared a very similar piece of hers a few months back, A 600-Year-Old Blueprint for Weathering Climate Change. That one at The Atlantic seems to be broken now, or their paywall is not working properly, so keep this one instead, and the topic is worth a revisit, even if you’d read the other back then.

Although most Americans today don’t know it, there were large centralized civilizations across much of North America in the 10th through 12th centuries. They built massive cities and grand irrigation projects across the continent. Twelfth-century Cahokia, on the banks of the Mississippi River, had a central city about the size of London at the time. […]

Leaders generally governed by facilitating decision-making in council meetings and public gatherings. They gave gifts to encourage cooperation. They heard disputes between neighbors over land and resources and helped to resolve them. Power and prestige came to lie not in amassing wealth but in assuring that the wealth was shared wisely. Leaders earned support in part by being good providers.

§ With AI, the future of Augmented Reality is in your ears. This interview with Dennis Crowley (Foursquare) is only so-so (sorry guys) but I think the premise of the title is correct and encourage you to keep an eye on it. “Then as you walk around the city, it will tell you things about certain places. Eventually, if I walk by a place where my friend was, it tells me that Alex was here two days ago. If I walk by a place and someone's inside, it tells me that Max is inside that place. A lot of it is still under development. Eventually, people leave a comment at a place. Imagine Twitter. It is as if you leave a tweet and you stick it in the ground. When you walk over it, you hear it.” I just republished this old Dispatch, scroll to “Hearing & AR” for some similar ponderings from various people.

Futures, Fictions & Fabulations

• Why every company needs a futurist-in-residence. “Identifying and tracking signals of change is the primary activity of any foresight work. It’s focused on spotting glimpses of the future that exist in the world today. A signal might be a new idea, behavior, technology, social context, or business model that may seem insignificant now, but could redefine industries or the world at large if it scales.” (I’m available for fractional roles, btw.)
• Speculative materialism: Redefining our relationship with materials. “Speculative Materialism invites us to rethink our relationship with materials, demonstrating how creative thinking and design innovation can mitigate contemporary climate challenges.”
• Futures toolkit for policymakers and analysts. It’s been around since 2014 but this is the recently published 2024 version. “The Futures Toolkit provides a set of tools to help you develop policies and strategies that are robust in the face of an uncertain future.”

Algorithms, Automations & Augmentations

• AI isn’t about unleashing our imaginations, it’s about outsourcing them. The real purpose is profit. “Artificial intelligence doesn’t just incrementally erode the rights of authors and other creators. These technologies are designed to replace creative workers altogether”
• Generative AI taught a robot dog to scramble around a new environment. “Now there’s a potentially better option: a new system that uses generative AI models in conjunction with a physics simulator to develop virtual training grounds that more accurately mirror the physical world.”
• Some of Substack’s biggest newsletters rely on AI writing tools. “A new analysis of Substack’s top newsletters estimated that around 10 percent publish AI-generated or AI-assisted content.”

Built, Biosphere & Breakthroughs

• Bringing life to L.A.’s infrastructure. “Living infrastructure is an invitation to honor place, grow participation, embody justice, foster resilience, and regenerate life.”
• 6 Questions you asked yourself about solar. “How fast will it take over, how fast are costs shrinking, why is it so cheap, what industries will it birth, how much surface will it take up, where will it appear first?”
• Lochaber’s Skipinnish Oak wins UK tree of the year. “Native woodland experts had no idea the tree existed until a gathering in 2009. It is thought to be at least 400 years old, and potentially up to 1,000 years old.”

Asides

• Gamer role introduced in children’s hospital. “A Scottish children’s hospital charity has introduced a gamer in residence for young patients in Glasgow. The new job involves visiting children to play video games with them, preventing boredom and providing some light relief.”
• Lidar mapping reveals mountainous medieval cities along the Silk Road. “When you were a medieval merchant leading a caravan carrying silk, perfumes, mirrors, and other goods you just bought in China, you had roughly a one-year-long journey ahead of you before you could sell them in Europe. The first really tough part of this journey was crossing the Tien Shan range using the mountain paths that avoided the Taklamakan Desert.”
• Port handwritten covers. “As well as a complete layout overhaul, Port’s recent redesign introduced an intriguing approach to its covers. Instead of a constant logo, the magazine title is handwritten by the issue’s cover star.”