Most digital art, even when it calls itself generative, still behaves like photography.

An algorithm runs. A result is rendered. The output is fixed, preserved, circulated. The process may be elegant or complex, but it resolves. What remains is an image... a trace of computation rather than computation itself.

This model has served digital art well, but it quietly inherits assumptions from older media: that artworks finish, that time is external rather than internal, and that preservation means keeping an object unchanged.

Yet for decades, artists and theorists have been circling a different question: 𝒘𝒉𝒂𝒕 𝒉𝒂𝒑𝒑𝒆𝒏𝒔 𝒘𝒉𝒆𝒏 𝒂𝒏 𝒂𝒓𝒕𝒘𝒐𝒓𝒌 𝒊𝒔 𝒏𝒐𝒕 𝒂𝒏 𝒐𝒃𝒋𝒆𝒄𝒕, 𝒃𝒖𝒕 𝒂 𝒔𝒚𝒔𝒕𝒆𝒎? 𝑾𝒉𝒂𝒕 𝒉𝒂𝒑𝒑𝒆𝒏𝒔 𝒘𝒉𝒆𝒏 𝒕𝒊𝒎𝒆 𝒊𝒔 𝒏𝒐𝒕 𝒂 𝒕𝒉𝒆𝒎𝒆 𝒃𝒖𝒕 𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆, 𝒘𝒉𝒆𝒏 𝒄𝒉𝒂𝒏𝒈𝒆 𝒊𝒔 𝒏𝒐𝒕 𝒅𝒆𝒄𝒂𝒚 𝒃𝒖𝒕 𝒃𝒆𝒉𝒂𝒗𝒊𝒐𝒖𝒓?

𝑼𝒏𝒇𝒊𝒏𝒊𝒔𝒉𝒆𝒅 𝒒𝒖𝒆𝒔𝒕𝒊𝒐𝒏𝒔 𝒊𝒏 𝒔𝒚𝒔𝒕𝒆𝒎𝒔 𝒂𝒏𝒅 𝒄𝒐𝒏𝒄𝒆𝒑𝒕𝒖𝒂𝒍 𝒂𝒓𝒕

In the 1960s and 70s, systems aesthetics and conceptual art shifted attention away from surface form toward process, rules and relationships. Sol LeWitt’s assertion that “the idea becomes a machine that makes the art” foregrounded generative logic over produced artifacts. Hans Haacke treated artworks as dynamic systems embedded in environmental, social or institutional feedback loops. Others worked with scores, instructions, and serial structures (works designed to unfold rather than resolve).

These approaches challenged the artwork as a stable object, but they faced a practical limit. Systems required caretakers. Scores required performers. Software art required servers, institutions or proprietary platforms to keep running. When those supports disappeared, the work froze into documentation or vanished altogether.

The system remained conceptually central, but materially fragile.

𝑻𝒉𝒆 𝒎𝒊𝒔𝒔𝒊𝒏𝒈 𝒂𝒇𝒇𝒐𝒓𝒅𝒂𝒏𝒄𝒆: 𝒑𝒆𝒓𝒔𝒊𝒔𝒕𝒆𝒏𝒕 𝒄𝒐𝒎𝒑𝒖𝒕𝒂𝒕𝒊𝒐𝒏

What these practices lacked was not imagination, but infrastructure.

A long-running artwork requires a substrate that is always on, publicly accessible, and resistant to institutional failure. It needs determinism without central control, persistence without ownership, and the ability to accept interaction without collapsing into chaos. Historically, that combination did not exist.

Most computational environments are contingent. Servers go offline. Platforms deprecate APIs. File formats rot. Even well funded institutions struggle to preserve living software for decades, let alone centuries.

This is where blockchains matter... not as markets or ideologies, but as runtime environments.

A public blockchain is an unusual thing: a globally replicated computer that executes the same program, in the same way, indefinitely, without a central operator. It is slow, constrained, and expensive... but it is persistent. Once a program is deployed, it cannot be quietly altered or shut down. It continues to run as long as the network exists.

That persistence changes what kinds of art are possible.

𝑻𝒆𝒓𝒓𝒂𝒇𝒐𝒓𝒎𝒔: 𝒂 𝒄𝒂𝒔𝒆 𝒔𝒕𝒖𝒅𝒚 𝒊𝒏 𝒓𝒖𝒏𝒕𝒊𝒎𝒆 𝒂𝒓𝒕

Terraforms is a long running generative artwork deployed on Ethereum in late 2021. At a surface level, it appears as 9,910 animated, text-based parcels rendered from Unicode characters, collectively forming a three-dimensional structure known as the Hypercastle.

Each parcel presents a 32×32 grid of characters forming an evolving topography. Colours cycle. Elevations oscillate. The visuals are generated entirely on chain, using deterministic inputs such as blockchain state to animate continuously.

This surface description is deliberately incomplete.

Terraforms is not a collection of images. It is a continuously running program, and its images are merely views into a shared simulation. The artwork does not exist at any single resolution or representation; it exists as code in execution.

The Hypercastle (the full structure composed of all parcels) is best understood as a single artwork instantiated across many coordinates. Individual parcels have meaning only in relation to the whole and to each other. Position, elevation, and behavior are products of a generative field that never stops computing.

This is why Terraforms is often described by its creator, 0x113d, as runtime art. The work was not generated and stored; it continues to run.

𝑺𝒕𝒂𝒕𝒆, 𝒊𝒏𝒕𝒆𝒓𝒗𝒆𝒏𝒕𝒊𝒐𝒏, 𝒂𝒏𝒅 𝒅𝒆𝒄𝒂𝒚

One of Terraforms’ most consequential features is that it allows intervention.

Parcels exist in distinct modes. In its default state, a parcel displays generative terrain. Parcels can enter a Daydream state, enabling Unicode glyph drawings to be written on-chain, permanently altering the system’s state.

This action is not reversible. Once a parcel leaves its original terrain state, it cannot return. The pristine landscape is overwritten.

At the scale of the Hypercastle, these interventions matter. The system incorporates long-term decay mechanics: over millennia, the structure erodes unless enough parcels remain in active, dream like states. Preservation, paradoxically, requires participation. Non-intervention is not neutral.

This formalises a tension digital art has rarely been forced to confront: permanence versus care. Terraforms stages impermanence inside a supposedly immutable medium, compelling participants to choose between conservation and expression.

In doing so, it echoes land art’s engagement with entropy and time, transposed into a computational register.

𝑫𝒆𝒍𝒆𝒈𝒂𝒕𝒊𝒐𝒏 𝒂𝒏𝒅 𝒎𝒖𝒍𝒕𝒊 𝒂𝒈𝒆𝒏𝒕 𝒂𝒖𝒕𝒉𝒐𝒓𝒔𝒉𝒊𝒑

Terraforms also destabilises authorship.

Owners can delegate the ability to modify their parcels to other individuals or even to smart contracts without transferring ownership. Drawings, transformations, and behaviours can be authored collectively or programmatically.

Recent upgrades extend this logic through broadcast mechanisms. Parcels can function as antennas, receiving generative signals from external contracts and exhibiting new behaviours in response. The artwork becomes a network of interacting programs rather than a closed system.

Authorship here is neither singular nor final. It is distributed across time, agents and code.

𝑾𝒉𝒚 𝒕𝒉𝒊𝒔 𝒎𝒂𝒕𝒕𝒆𝒓𝒔 𝒃𝒆𝒚𝒐𝒏𝒅 𝒄𝒓𝒚𝒑𝒕𝒐

Terraforms is often misunderstood because it is approached through the wrong lens. Seen as a speculative asset or a static NFT collection, it appears opaque or overengineered. This misreading reflects prevailing assumptions more than the work itself.

Historically, significant artworks are often those that redefine a medium’s affordances rather than refine its aesthetics. Terraforms demonstrates that blockchains can host long-duration, stateful, interactive artworks. Not merely store representations of them.

It reopens questions digital art has asked for decades but lacked the tools to fully pursue: How do you design art for centuries? What does preservation mean when change is intrinsic? How do systems carry culture, not just content?

In this sense, Terraforms is less a product than a proposition. It offers a blueprint for treating computation itself as a site of artistic inquiry... persistent, participatory, and irreducible to any single view.

𝑨𝒇𝒕𝒆𝒓 𝒓𝒆𝒔𝒐𝒍𝒖𝒕𝒊𝒐𝒏

Some artworks end when they are completed. Others end when they are documented. Terraforms does neither.

It continues to compute, to decay, to accept intervention and to change. Not as metaphor, but as behaviour. Its significance lies not in novelty or scale, but in its insistence that art on a networked computer need not resolve at all.

As more artists begin to treat blockchains not as storage layers but as expressive machines, Terraforms may come to be seen less as an anomaly and more as an early, durable answer to a long standing question:

𝑾𝒉𝒂𝒕 𝒅𝒐𝒆𝒔 𝒂𝒏 𝒂𝒓𝒕𝒘𝒐𝒓𝒌 𝒅𝒐 𝒘𝒉𝒆𝒏 𝒊𝒕 𝒏𝒆𝒗𝒆𝒓 𝒇𝒊𝒏𝒊𝒔𝒉𝒆𝒔?

Most digital art, even when it calls itself generative, still behaves like photography.

An algorithm runs. A result is rendered. The output is fixed, preserved, circulated. The process may be elegant or complex, but it resolves. What remains is an image... a trace of computation rather than computation itself.

This model has served digital art well, but it quietly inherits assumptions from older media: that artworks finish, that time is external rather than internal, and that preservation means keeping an object unchanged.

Yet for decades, artists and theorists have been circling a different question: 𝒘𝒉𝒂𝒕 𝒉𝒂𝒑𝒑𝒆𝒏𝒔 𝒘𝒉𝒆𝒏 𝒂𝒏 𝒂𝒓𝒕𝒘𝒐𝒓𝒌 𝒊𝒔 𝒏𝒐𝒕 𝒂𝒏 𝒐𝒃𝒋𝒆𝒄𝒕, 𝒃𝒖𝒕 𝒂 𝒔𝒚𝒔𝒕𝒆𝒎? 𝑾𝒉𝒂𝒕 𝒉𝒂𝒑𝒑𝒆𝒏𝒔 𝒘𝒉𝒆𝒏 𝒕𝒊𝒎𝒆 𝒊𝒔 𝒏𝒐𝒕 𝒂 𝒕𝒉𝒆𝒎𝒆 𝒃𝒖𝒕 𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆, 𝒘𝒉𝒆𝒏 𝒄𝒉𝒂𝒏𝒈𝒆 𝒊𝒔 𝒏𝒐𝒕 𝒅𝒆𝒄𝒂𝒚 𝒃𝒖𝒕 𝒃𝒆𝒉𝒂𝒗𝒊𝒐𝒖𝒓?

𝑼𝒏𝒇𝒊𝒏𝒊𝒔𝒉𝒆𝒅 𝒒𝒖𝒆𝒔𝒕𝒊𝒐𝒏𝒔 𝒊𝒏 𝒔𝒚𝒔𝒕𝒆𝒎𝒔 𝒂𝒏𝒅 𝒄𝒐𝒏𝒄𝒆𝒑𝒕𝒖𝒂𝒍 𝒂𝒓𝒕

In the 1960s and 70s, systems aesthetics and conceptual art shifted attention away from surface form toward process, rules and relationships. Sol LeWitt’s assertion that “the idea becomes a machine that makes the art” foregrounded generative logic over produced artifacts. Hans Haacke treated artworks as dynamic systems embedded in environmental, social or institutional feedback loops. Others worked with scores, instructions, and serial structures (works designed to unfold rather than resolve).

These approaches challenged the artwork as a stable object, but they faced a practical limit. Systems required caretakers. Scores required performers. Software art required servers, institutions or proprietary platforms to keep running. When those supports disappeared, the work froze into documentation or vanished altogether.

The system remained conceptually central, but materially fragile.

𝑻𝒉𝒆 𝒎𝒊𝒔𝒔𝒊𝒏𝒈 𝒂𝒇𝒇𝒐𝒓𝒅𝒂𝒏𝒄𝒆: 𝒑𝒆𝒓𝒔𝒊𝒔𝒕𝒆𝒏𝒕 𝒄𝒐𝒎𝒑𝒖𝒕𝒂𝒕𝒊𝒐𝒏

What these practices lacked was not imagination, but infrastructure.

A long-running artwork requires a substrate that is always on, publicly accessible, and resistant to institutional failure. It needs determinism without central control, persistence without ownership, and the ability to accept interaction without collapsing into chaos. Historically, that combination did not exist.

Most computational environments are contingent. Servers go offline. Platforms deprecate APIs. File formats rot. Even well funded institutions struggle to preserve living software for decades, let alone centuries.

This is where blockchains matter... not as markets or ideologies, but as runtime environments.

A public blockchain is an unusual thing: a globally replicated computer that executes the same program, in the same way, indefinitely, without a central operator. It is slow, constrained, and expensive... but it is persistent. Once a program is deployed, it cannot be quietly altered or shut down. It continues to run as long as the network exists.

That persistence changes what kinds of art are possible.

𝑻𝒆𝒓𝒓𝒂𝒇𝒐𝒓𝒎𝒔: 𝒂 𝒄𝒂𝒔𝒆 𝒔𝒕𝒖𝒅𝒚 𝒊𝒏 𝒓𝒖𝒏𝒕𝒊𝒎𝒆 𝒂𝒓𝒕

Terraforms is a long running generative artwork deployed on Ethereum in late 2021. At a surface level, it appears as 9,910 animated, text-based parcels rendered from Unicode characters, collectively forming a three-dimensional structure known as the Hypercastle.

Each parcel presents a 32×32 grid of characters forming an evolving topography. Colours cycle. Elevations oscillate. The visuals are generated entirely on chain, using deterministic inputs such as blockchain state to animate continuously.

This surface description is deliberately incomplete.

Terraforms is not a collection of images. It is a continuously running program, and its images are merely views into a shared simulation. The artwork does not exist at any single resolution or representation; it exists as code in execution.

The Hypercastle (the full structure composed of all parcels) is best understood as a single artwork instantiated across many coordinates. Individual parcels have meaning only in relation to the whole and to each other. Position, elevation, and behavior are products of a generative field that never stops computing.

This is why Terraforms is often described by its creator, 0x113d, as runtime art. The work was not generated and stored; it continues to run.

𝑺𝒕𝒂𝒕𝒆, 𝒊𝒏𝒕𝒆𝒓𝒗𝒆𝒏𝒕𝒊𝒐𝒏, 𝒂𝒏𝒅 𝒅𝒆𝒄𝒂𝒚

One of Terraforms’ most consequential features is that it allows intervention.

Parcels exist in distinct modes. In its default state, a parcel displays generative terrain. Parcels can enter a Daydream state, enabling Unicode glyph drawings to be written on-chain, permanently altering the system’s state.

This action is not reversible. Once a parcel leaves its original terrain state, it cannot return. The pristine landscape is overwritten.

At the scale of the Hypercastle, these interventions matter. The system incorporates long-term decay mechanics: over millennia, the structure erodes unless enough parcels remain in active, dream like states. Preservation, paradoxically, requires participation. Non-intervention is not neutral.

This formalises a tension digital art has rarely been forced to confront: permanence versus care. Terraforms stages impermanence inside a supposedly immutable medium, compelling participants to choose between conservation and expression.

In doing so, it echoes land art’s engagement with entropy and time, transposed into a computational register.

𝑫𝒆𝒍𝒆𝒈𝒂𝒕𝒊𝒐𝒏 𝒂𝒏𝒅 𝒎𝒖𝒍𝒕𝒊 𝒂𝒈𝒆𝒏𝒕 𝒂𝒖𝒕𝒉𝒐𝒓𝒔𝒉𝒊𝒑

Terraforms also destabilises authorship.

Owners can delegate the ability to modify their parcels to other individuals or even to smart contracts without transferring ownership. Drawings, transformations, and behaviours can be authored collectively or programmatically.

Recent upgrades extend this logic through broadcast mechanisms. Parcels can function as antennas, receiving generative signals from external contracts and exhibiting new behaviours in response. The artwork becomes a network of interacting programs rather than a closed system.

Authorship here is neither singular nor final. It is distributed across time, agents and code.

𝑾𝒉𝒚 𝒕𝒉𝒊𝒔 𝒎𝒂𝒕𝒕𝒆𝒓𝒔 𝒃𝒆𝒚𝒐𝒏𝒅 𝒄𝒓𝒚𝒑𝒕𝒐

Terraforms is often misunderstood because it is approached through the wrong lens. Seen as a speculative asset or a static NFT collection, it appears opaque or overengineered. This misreading reflects prevailing assumptions more than the work itself.

Historically, significant artworks are often those that redefine a medium’s affordances rather than refine its aesthetics. Terraforms demonstrates that blockchains can host long-duration, stateful, interactive artworks. Not merely store representations of them.

It reopens questions digital art has asked for decades but lacked the tools to fully pursue: How do you design art for centuries? What does preservation mean when change is intrinsic? How do systems carry culture, not just content?

In this sense, Terraforms is less a product than a proposition. It offers a blueprint for treating computation itself as a site of artistic inquiry... persistent, participatory, and irreducible to any single view.

𝑨𝒇𝒕𝒆𝒓 𝒓𝒆𝒔𝒐𝒍𝒖𝒕𝒊𝒐𝒏

Some artworks end when they are completed. Others end when they are documented. Terraforms does neither.

It continues to compute, to decay, to accept intervention and to change. Not as metaphor, but as behaviour. Its significance lies not in novelty or scale, but in its insistence that art on a networked computer need not resolve at all.

As more artists begin to treat blockchains not as storage layers but as expressive machines, Terraforms may come to be seen less as an anomaly and more as an early, durable answer to a long standing question:

𝑾𝒉𝒂𝒕 𝒅𝒐𝒆𝒔 𝒂𝒏 𝒂𝒓𝒕𝒘𝒐𝒓𝒌 𝒅𝒐 𝒘𝒉𝒆𝒏 𝒊𝒕 𝒏𝒆𝒗𝒆𝒓 𝒇𝒊𝒏𝒊𝒔𝒉𝒆𝒔?