Living Ideas – the organic future
Designers of the Future, LLMs and Pottery Making, The Interface Revolution
TLDR Version
👋 Hi frens!
My rational and analytical mind says I should start by sharing my motivations for writing this piece. It’s the boldest of all the articles so far for me. I can’t precisely track “the one” inspiration source for writing it. Writing this is something different, something else.
I’m writing this partly to capture my point of view on the recent parabolic ascent of AI technology (whether or not this is hype is irrelevant). I wanted to focus on the upside and cut the chains of my inner techno-optimist dog to see where he would roam when cut loose. I sincerely hope he won’t do much harm along the way in lieu of the AI alignment challenge, one of this century's most critical challenges. It can make us or break us forever.
This piece is also inspired by Packy's McCormick's experiment with writing fiction titled "Idea Olympics". What you will read in my essay is a continuation and deeper exploration of Packys' vision. Packy, thank you for being the evergreen source of optimism.
Researching for this article was a blast for me, and I would love to write such, for the lack of a better word, “weird” things occasionally 🙂 Please leave feedback and let me know if you would like to see more of such explorations in the future.
Ok, enough rambling. Let’s jump in!
Wherever I gaze, ideas shine bright.
Ripe and red from absorption of light.
Endless rows going into the night,
Which one of you will serve me right?
The data their water,
and the code is their form,
curiosity their sunlight,
Behold - how they grow!
~ Jon The Skin
This essay explores tangible ideas and their relationship with Artificial Intelligence. It’s an attempt to paint a vision for the future, consider what the future could feel like, and what sensible strategies can help us get there. During this two-act journey, we will explore how future ideas will be created, discovered, browsed, purchased, and used.
Living Idea Concept
Before we enter the exciting realm of the Living Ideas – let’s define what they are. So what’s an idea? Let’s quickly look at the word's origin – from the Greek idein – “to see” or idea – “form, pattern”. An idea is a formulated thought that has a distinct pattern. These thoughts can be about many things, persons, places, or concepts.
Ideas have a wide array of different characteristics. One of these characteristics is their “liveliness”. I want to explore this topic deeper and look at ideas possessing this unique quality. I call these ideas – Living Ideas or LIs (read “el-ais”).
The Living Ideas are the ideas that one can somehow interact with before actually adopting them – think “testing the product before purchase, “try before you buy” way. The second point about Living Ideas is that they change when exposed to other ideas, enrich themselves with every interaction with the environment, and are not stable in the sense of being “fixed things”. Their patterns morph and change, sometimes a little, other times a lot. The third aspect of Living Ideas is that they can come to being only when sufficiently enabled by the AI. The fourth thing to remember about the Living Ideas is that they occupy a shared and organic space.
Where do the LIs live
The Living Ideas discussed in this essay need some space to live and exist in, right? I think that the first LIs will live in cyberspace exclusively. What do I mean by that? This space has many names. Some call it the metaverse, others multiverse or omniverse (Nvidia). Regardless of how we call it today and how its exact definitions will change in the future, one thing is true – these different intangible spaces are here to stay.
Living Ideas (LIs) will bleed into our emerging digital spaces as more of our lives are led in the digital. Notice a clear trend here: As our lives become increasingly digitized, what we consider a “digital” separate from “reality” merges into the “digital realities”. The metaverse concept is drawing big players and bringing high stakes to the table. Facebook/Meta wants to expand and take control over the Metaverse from many different angles. Oculus + subsidizing users move into the Metaverse are just a few examples. Around 2021-2022, Meta ran out from a pool of “throat share”. They got over the limit of their Total Addressable Market (TAM) and must expand their offering to higher-level markets to continue making investors happy. Their actions are now the consequences of a company reaching a limit to the capital it can extract from societies’ present relationships with technology. Since not much of our lives are yet platformized enough – Meta wants to use this and expand activities available for people. And the expansion direction? Yep – the digital, the metaverse, the boundless space. In the same space in which LIs will exist.
Use cases for LIs
So, what are the use cases for these LIs? In other words, to what needs would these ideas respond? I think LIs can do many things – but I want to focus on four potential applications: solve, build, blend, and simulate.
Looking at LIs through the lenses of potential applications will better clarify what LIs are. For example, for software products space – LIs will take the shape of ideas that solve concrete user problems. Certain ideas that help build game worlds for players to traverse and explore will emerge in the gaming industry. LIs of this type will be used primarily for building the digital-first reality. These game worlds will exist in digital spaces that will begin to blend with our reality. For the Metaverse, LIs will optimize blending the digital with the real. Last but not least, a distinct group of LIs will emerge that focuses on simulating reality – a gateway for this is already emerging through the increased use of Digital Twins (Nvidia, Microsoft, Autodesk, Bosch, IBM, Siemens, you get the point).
LIs for building
Let’s look at the “building” use cases for Living Ideas. To do that, we will visit the gaming industry. Games are, after all – big systems of ideas nested within smaller ideas. The game developers create and build ideas every day.
Every game has a world – a totality of Environments, Architecture, Rules, Characters, Items, and Interactions that make up the player's experience. Today, all the elements that make up the game world are ideated, refined, and put into the game by artists and developers. Creating such assets is undoubtedly time-consuming and involves multiple interactions with different specialists and a lot of trial and error. Yet, such is the nature of ideas. Some of them are complicated.
We must go one step deeper to illustrate the true nature of LIs better.
Take any game where some characters appear (NPCs or not). Now let’s zoom in on our character – We can call him “Link”. Link is a sword-fighting warrior. He looks a certain way, wears dark clothes, and speaks funnily.
But Link is also something more than the sum of its parts. He’s an entire, coherent idea that the player can interact with. He has been created by someone, produced by a specific development process. Here’s the crucial insight – a game character production is, in a way, an idea production. Sure, it is a limited idea wrapped in some visual concept and code – but it is an idea nevertheless.
How is this related to applying Living Ideas for building stuff? I think the true potential of future Living Ideas will shine to the extent that they can coherently build up and capture the entirety of what our Link is. Let me explain what I mean.
When artists conceptualize such characters and want to bring them to life – they must consider a myriad of elements:
Game engine (possibilities and limitations)
Character look and feel (model)
Movement (rig)
Behavior (scripts)
Voice (voice and tone)
Lore
Backstory
Role in the story
Relationships with the world
Equipment, items
Attributes, quirks, etc.
Current technology for bringing Link “to life” is quite convoluted and complicated. The game character production pipeline has a lot of moving parts. With current technology, joining/integrating these parts takes time. The emergence of Living Ideas faces many barriers, including time constraints. Yet, the character production pipeline is bound to be disrupted by AI. A similar disruption will happen to the process of producing ideas.
With the advent of AI and the right LLMs and tools built around them, character production (idea production) and many other building processes will become hugely simplified and automatized. You don’t need to look far for examples. Game developers can already use Nvidia ACE for games to customize and deploy LLMs through the cloud or PC to generate intelligent NPCs. Product researchers in the early discovery stages no longer need to recruit users to kickstart their work. They can use the virtual ones that AI has synthesized.
Let’s get back to our character design.
It would seem that having so many moving parts that constitute just one character (sic!) would lead to the natural emergence of liveliness to this “idea”. Yet having something built from many parts doesn’t necessarily mean it’s alive. Think of the animal cell being more than the sum of its parts – alive-like. With the current state of being, the game characters only get “sort of alive” once put into a world they were designed to inhabit. This world they exist in needs careful and time-consuming building as well! Inversely, without the surrounding context, ideas get bleak, disjointed, and separate. They just can’t stand on their own. They need the context.
I think the Living Ideas we are trying to devise here will be the first of the “new breed”. Meaning they will be able to stand on their own. Regardless of the context or platform. How? The short answer is that the necessary context will be provided live on-demand with the help of sufficiently advanced and trained LLMs. The long answer? They will require a convergence of multiple threads in technology, design, interface, and human behavior. Let’s pull on some of these threads.
Ideas on their own
Picture this: an idea contained in a sphere submerged in green goo. The strange goo not only provides all the necessary nutrients to the idea but also serves the role of a context-providing virtual environment through which the idea can relate to any other idea.
I believe that integrating all previously listed game character elements into coherent, easily transferable, and widely applicable sphere-like singular idea is the future. This is how the next wave of builders will approach the creation in the digital world. The AI will “fill in the blanks” between the previously disjointed elements like art, rigging, backstory, and behavior. Plus, most importantly – it will provide context. A combination of LLMs and tech will animate the idea. The ghost will inhabit the shell. And that’s how the LIs will be born.
Imagine your favorite game, movie, or book characters as living things you can interact with in real-time. Characters with their DNA that’s true to the canon or perhaps slightly modified to your heart's content – pick your favorite. This will be the first time you actually get to pick your favorite because the character is… a Living Idea! A personalized, context–full and coherent thing. That’s being shaped through your interactions with it, ever-changing, ever-evolving. No longer static and bound to only one realm – the character comes alive right before your eyes.
Ideas of the future
When conceptualizing LIs, I considered how they would manifest daily. I have asked myself: “What will constitute the most powerful ideas in the future?”. Powerful is understood as being able to touch as many people as possible meaningfully.
With intuitions' help, product-building experience, a bit of philosophy, and an undying urge to organize reality into parts that make sense – I think I have arrived at a convincing formula.
The most powerful ideas in the future will be Designed, Distributed and Distinct. Let me explain why.
Let’s backtrack and consider how we could assess the Living Ideas. There are seemingly infinite ways to do it, depending on what you believe is valuable and meaningful. I simplified these philosophical considerations to the most basic and practical themes. When assessing the quality of LIs we should look at three general criteria:
Interdependence – The extent to which the LI is separate from other LIs
Desirability – How valuable is a given LI for people
Tangibility – LI connectedness with the reality
Within each of these three criteria, there’s a spectrum. The most powerful ideas will subtly balance the two far sides of the spectrum. Let’s take a look at what’s on these spectrums.
Interdependence
Spectrum elements for Interdependence reflect how ideas relate to each other. They range from a) standing alone to b) being deeply enmeshed within a web of different ideas. The most potent ideas will sit close to the middle of the spectrum, balancing the tradeoffs. Let’s go from the far left and look at the Autonomous characteristics.
Autonomous: These are LIs that can stand independently without the support of other ideas. They are complete and self-sufficient, often providing a single solution to a specific problem. This is similar to a tool with one function that performs exceptionally well.
Why do I think that entirely autonomous ideas won’t take off? They lose the supremacy battle because they are too separate from the other ideas. This separation means they can’t benefit from the surging’ surging popularity and positive influence. Idea cross-pollination is challenging, and stagnation becomes likely.
Think of Japan, which isolated itself from the rest of the world for 265 years (XVII-XIX). Sure, they lived in peace and considerable prosperity. The arts and culture flourished. Yet, it was all achieved at a significant cost – the cost of technological, economic, and intellectual stagnation. When Japan did open up, it found itself significantly behind other world powers, necessitating rapid and often difficult modernization and Westernization efforts during the Meiji Restoration.
Distributed (the sweet spot): This position in the spectrum refers to LIs crafted not only to operate harmoniously within a given system but also to form a dynamic network with a range of other ideas, often spanning multiple domains. Synergistic ideas offer cumulative value, enhancing their utility when they engage with one another.
These ideas are akin to apps operating within the same software environment, enhancing each other's capabilities. They are also similar to nodes within a blockchain or the internet, where each node's contribution is vital to the strength and capability of the entire network.
Overextended: LIs become convoluted through excessive interdependence, resulting in complexity that obstructs usefulness and clarity. At this point in the spectrum, the ideas may be so interwoven that they create a fragile ecosystem where the failure or confusion of one idea can cascade and disrupt the entire system. Overextension can lead to inefficiency, lack of focus, and difficulty in understanding or utilizing ideas effectively. It is the conceptual equivalent of a “too many cooks in the kitchen” scenario, where the integration of ideas becomes overcomplicated and counterproductive. We don’t want that.
Let’s move on to the Desirability criteria now.
Desirability
Spectrum elements for desirability reflect how valuable the ideas are to people.
Useful: This element refers to ideas that are primarily valued for their utility and ability to perform a specific function efficiently. The focus is on practical application, providing straightforward and effective solutions to users' needs – a function over form. Think of a spoon or a hammer.
Usefulness is necessary but not sufficient for the best Living Ideas. There’s great competition and commodification at the mere usefulness level. And I think that this will continue in the future. Truly Living Ideas will not only integrate the “plain” usefulness but also the unique experience.
Designed (the sweet spot): Ideas in this category embody a harmonious blend of form and function, where aesthetic appeal is just as important as practicality. These are well-considered creations that not only serve their purpose but also deliver a pleasing user experience through meticulous design. Think Apple products.
Artistic: At this level, ideas transcend utility and enter the realm of expression, evoking emotional responses and thought-provoking interactions. Artistic ideas are appreciated for their ability to inspire, challenge perceptions, and provide a profound aesthetic or conceptual impact beyond conventional use. They focus on conveying a message and making us feel certain emotions. Think of famous art installations.
Let’s look at the third and last criterion – Tangibility.
Tangibility
Spectrum elements for tangibility reflect how much in touch with the reality the LI is.
Based: These ideas are entirely based on reality. Ideas in this category are grounded in the physical world and adhere strictly to empirical evidence and actual experiences. They prioritize factual and tangible outcomes over theoretical or speculative concepts, ensuring a solid, real-world foundation for their application. Fuzzy = Bad. Think of everything driven by hard numbers, quantification, and precise computation as the primary function – i.e., Wolfram Alpha.
Distinct (the sweet spot): These ideas are “reality inspired”, but have a qualitative component. They draw inspiration from reality but interpret it through a unique lens, offering a fresh perspective or novel approach. While they are still connected to the tangible world, they are characterized by a certain level of creativity, interpretation, and innovation that sets them apart.
Think 3D Printing in medicine or Kickstarter. For example, Kickstarter transforms traditional funding into a distinctive crowdsourcing platform, realizing innovative, reality-inspired projects that defy conventional financing. These campaigns are not inventing the categories anew; they're reimagining them in ways that resonate uniquely with potential backers.
Intangible/Abstract: This end of the spectrum encompasses ideas that exist primarily in the realm of thought, theory, or imagination with no direct physical manifestation. These ideas are often conceptual, dealing with notions, possibilities, or visions that challenge the conventional boundaries of physical existence. Think about the nature of reality, the meaning of life, or mindfulness.
The only thing that’s hard about the intangible and abstract is that they are hard to grasp 😅. Humans don’t intentionally contend with them regularly. Thus, I believe that highly abstract ideas are at a disadvantage when it comes to taking off as Living Ideas.
Alright, we looked into what qualities these Living Ideas have and which are likely to be the most powerful. Now let’s get to the fun part and answer the question – What would it look like to create a Living Idea?
Editing Ideas
The standard processes we know and use today for creating and building ideas must change. Whether these processes relate to creating a digital asset, a game NPC character, an in-game item, a synthetic persona, or a digital twin – pick your favorite, they will be different.
We will need new tools for that because the current tools are limited by nature. Why? They have been thought out and developed with a legacy mindset. The first important distinction is that the most powerful future ideas won’t be created. They will be edited. Edited – from a pulp of material that can be shaped into anything. Living Ideas of the future won’t have creators. They will have editors. I will explain why and how in a minute.
At this point, we need to remind ourselves that when talking about LIs, we talk about creating/editing the wildest and most powerful ideas there have ever been. It’s not an easy feat. It's quite overwhelming, actually. To remedy that, I believe that the exploration of new tools and techniques should first of all come from the simplicity perspective.
The editor perspective
Let’s imagine what it would look like to create/edit a Living Idea from an editor's perspective.
My intuition tells me that editing Living Ideas will look and feel like… making pottery. There will be the editor – You, the clay – which is the material, and some shaping device (in pottery making, it’s usually hands) – as the device that does the shaping, plus some base – to hold it all together while the editing process takes place.
I want to cover these elements in more detail because it's fascinating. Also, there will be an entire section about the editors' role later in the essay, so I will skip it now. Let’s start with quickly exploring what would be at the base of the shaping act.
Base
The base is straightforward. The base IS the LLMs that understand human speech and aggregate multiple declared and undeclared inputs (like images, sounds, etc.). Now, onto the Material.
Material
In pottery making, the “clay” has particular properties and different “liveliness” to it. Its properties impact the final result and the experience of shaping it. Editing LIs will require certain material “properties” as well. What you can create at the end is often only as good as the material from which it was made. This is true in pottery making and interacting with LLMs (Large Language Models). In AI circles, there’s this rule of thumb saying that output quality is proportional to input quality. And it’s very accurate. Provide a poor prompt for ChatGPT – and get a meek output. Improve the prompt and see the output quality significantly increase. Simple.
Now, let’s merge the pottery example with what we know about AI models. The material IS the Data. After all, these LLMs need quality data to perform well too. The better the data quality and the more data AI has, the better the context is inferred by AI. Concurrently, the better the context inference – the more the final creation resembles the creator's intent. There’s also an additional fascinating & paradoxical dynamic taking place here. It’s related to the user experience of the editor.
Editor experience
The material must allow for being shaped not only through the declared commands/prompts but also through information that the creator DID NOT declare explicitly. Now, look at this from the artisan perspective. The artisan who is in the state of creative flow simply wants to a) create and b) don’t be bothered about anything else. Many creatives will admit that there’s no room for “analytical stagger” in the act of creation. Analysis makes us lose our flow; if not managed properly, too much of it done in the wrong place and time will be detrimental to the creation process. In such an advanced artistic process as the creation/editing of Living Ideas, there is no time nor space to declare every single bit of information and imprint it onto the material. So, how can we resolve this paradox?
The editor needs AI ingenuity and inference capabilities to sort of “work it out” in the background. Again, we must appreciate the previously mentioned function of the AI, which is “filling in the blanks” for us.
The seamlessness of the entire editor experience is a function of the right combination of the declared AND the highly relevant undeclared information. Making it work is a tough nut to crack – worthy of an entirely separate essay.
Let’s leave the material considerations aside and explore how the editor would shape the material.
Shaping device
In pottery making, the artisan uses her hands to shape the material. In the context of editing LIs this obviously wouldn’t do. Hands are too much of a limiting and clunky interface for this complex job. The truth is that every interface will somehow slow down the editor. So, what would be the ultimate interface that’s within our reach? How would the editor shape LIs? Well, I think the Living Ideas of the future will be created/edited using… words.
Human speech will become the grand idea-shaping device of the future. The speech is the next “wonder interface” – an important step in our eventual transition to the end game, which is a direct brain-to-brain link (Neuralink). Creating new ideas will feel like making pottery with words. Editors will polish the idea by talking with it and the other ideas around it. I call it Creating through Discussing. And it’s already happening! The rapid evolution of LLMs towards more of a commodity service will enable more conversational styles of creating new stuff and putting it into the world. This is not science fiction – notice that programming already looks like that. Just look at Aleksandar Simovic experiments with conversational programming. Granted, the current state of conversational programming is a mere “prompt engineering” still, yet give it some time, and it will evolve into a “voice conversation” between an engineer and AI – how awesome is that?!
Even though speech will become the predominant shaping device for creating LIs – we will also observe the emergence of what I call “Editor Toolkits”. Why are tools necessary? Because they help us rein the properties of the material. Just as pottery makers use certain tools to control the result of their work better, the editors will have tools other than speech that help them better modulate the ideas (more on modulation later). I can think of at least five tool types.
Tools for Objects – primitives, basic form templates. Think Blender templates.
Tools for Concepts – concept manipulators (using Boolean operators). Think Figma or search queries.
Tools for Words – word transformer packages (sets of different word combinations). For example, Market + Dynamics = Market Dynamics.
Tools for Phenomena – physics system simulators. Think COMSOL or PhysX.
Tools for Output – for when the visual, audio, or other “perceivable” output is needed. Think huge media suite like Adobe Creative Suite.
These tools will function as plugins, enhancing the creators’ workflow and speeding up the process.
Shaping skills
For effective editing of Living Ideas, specific “shaping skills” are required in addition to appropriate tools that can harness the properties of the material.
Just as our ancestors made many different items across the ages – we too will need to get back to the basics of what it takes to create something. Inspired by the work of linguist Daniel Everett – I believe that to masterfully shape Living Ideas, one must use imagination, intention, planning, and memory.
What it means is that there won’t suddenly be fewer engineers or creators. Someone will still need to do the creative thinking for the machine. The specialist numbers will stay more or less constant, but what will change is the legacy tools these specialists use to create stuff. Think of all your programs like 3ds Max, Adobe, Figma, Audition, etc. All of their current forms will become obsolete. And the actual step-by-step processes for creating/editing stuff will change as well – more on that later.
Editing Process
My intuition is that a proposition of the step-by-step editing process should be simple or at least attempt to descend from the simplest form possible. Let’s take a closer look at the actual process.
Since speech will be the main interface, the process will begin with capturing the words that are spoken. The words will describe the thing the editor wants to bring to life. The editor will start wide and put a goalpost somewhere in the distance.
Then, in step two, the initial input will be treated by the so-called “modulators”. These modulators have one job. To adequately form, re-form, and de-form the initial input – without breaking it. Today, this role is fulfilled by different AI Agents that improve and assess a concept provided by human input.
The third stage is a refined idea that is a combination of initial prompt and modulator shaping that’s narrowed down to a single entity. Editors will need to go through multiple iterations of this three-step process, which will take the form of a dance.
At the end of the process – the Living Idea emerges, and the editor can finally cheer, “That’s exactly what I wanted!”. As a side note, it would be a cool exercise to dig deeper into each of these three steps, as multiple sub-processes are involved. Perhaps an idea for a future exploration?
Editor role
Everyone who shapes ideas with the help of AI becomes an editor. Those who will shape the Living Ideas in the future will be called “Scriptors” – since their craft relates to “inscribing” with words and shaping the digital clay.
The future Scriptor role shares many characteristics with the multidisciplinary designer of today. A good and highly sought-after Scriptor of the future should have two qualities.
On the one hand, she will have a “design soul”, and on the other hand, the openness & courage to go beyond any classification systems. The soul has this insatiable need to answer the question — "How things work?", and the radical openness guides the designer to explore domains that are foreign to her.
All this is guided by the desire to shape a beautiful answer to the first question. Through these dynamics, the Scriptor changes and adapts while remaining true to herself. Ever-changing and ever-evolving. They will also need to become systems thinkers that focus on systems design rather than design systems. The greatest Scriptors will get inspired by a wide array of different knowledge domains in their strive to keep up with the changing digital landscape. They will experiment with the Generative AI technology and race to create the most potent Living Ideas.
When thinking about the Scriptor role, we can observe similar dynamics for requirements for this role emerging across great prompters. A huge part of being a successful prompter boils down to having a wide symbolic vocabulary at your disposal. For example, a visual designer with arts education will have an easier time creating Living Ideas with the use of AI because she had to learn all these different art concepts, painting styles, grid structures, and visual communication styles used in culture across the ages.
Let’s reiterate the critical principle of working with GenAI – the end result is closely tied to the ability of the Scriptor to declare the intent accurately. So, as a Scriptor – actively working on expanding your vocabulary of concepts, classifications, and points of view will give you a wider pool of symbols to use in editing new Living Ideas. Contemporary designers will need to adopt the learner mindset and literally become sponges for new symbols. "Concept Vocabulary" will become the next highly sought-after soft skill of future creators.
Looking forward, the emergence of Living Ideas is a call for action to all visionaries, creatives, and technologists to reimagine our role in the act of creation. It challenges us to become curators of a digital ecosystem where ideas breathe, evolve, and interact autonomously. As we harness the tools and talents of this new era, we will devise new experiences that are as diverse as they are dynamic – bringing characters, concepts, and creations to life in ways we have only begun to imagine.
Stay tuned for Act II, where I will describe how the Living Ideas will be discovered, browsed, and purchased on the Grand Marketplace. We will also take a closer look at LIs architecture and how we will use them in practice.
Till the next time!
Thanks for reading
Jan
The inner techno-optimist dog ✨