AB5D

Fake Internet Money

Steve Pikelny · Project 152 · Factory

Edition size: 1,000 · First mint: 28 August 2021

Fake Internet Money #29 by Steve Pikelny, from the AB5D collection
Fake Internet Money #29 by Steve Pikelny. Copyright: CC BY-NC 4.0. From the AB5D collection.

Work and context

Fake Internet Money makes a banknote convincing enough to invite belief, then supplies reasons to withdraw it. Denominations, rosettes, signatures and serial numbers suggest an issuing authority. Yet the title announces a fiction, and the project statement refuses the promises normally attached to money. Steve Pikelny turns the gap between a recognizable financial object and an agreed value into the subject of a generative collection.1

Art Blocks identifies the series as a Factory release of 1,000 works, publicly released on 1 September 2021. The AB5D-held example is #29, whose owner matches an address in the current linked-wallet profile. Its indexed features call it Crypto, Classic and Official, while also recording a missing-pigmentation misprint. Those categories already reveal the collection's deliberate coexistence of authority and failure.123

The word Official needs particular care. It is a feature of the artwork, not an endorsement of its monetary status. A token can be an authentic member of this collection while depicting money that has no redemption promise. Likewise, the visible denomination belongs to the generated note; it does not establish the token's exchange value.14

Conception

The artist's project website links to a CryptoGateway dialogue dated 28 August 2021. It presents Pikelny discussing his interest in collecting banknotes, their intricate patterns and their uncertain position between currency and collectible. The site is part of his fictional publishing practice. In a later Art Blocks interview, he explicitly describes publishing fake interviews with himself. The dialogue is therefore useful as an artist-authored project statement, not as independent journalism or corroboration of its historical claims.567

Within that staged conversation, the artist explains his interest in something that initially resembles money but gradually feels wrong. Altered proportions, absent familiar elements, ticket-like arrangements, saturated colours and simulated printing failures all pull the object away from its apparent model. He also connects the rosette experiments with I Saw It in a Dream, describing that earlier project's core pattern as an outcome of work toward this one.5

A real curator interview published by LaCollection in September 2022 makes the larger aim explicit. Pikelny says the shared rosette language gives the edition coherence while differences in denomination, serial numbers and misprints invite different valuations. He considers the collection as a whole the artwork, because its meaning depends on comparison among bills. The point is not a prediction of what any particular token should cost, but a question about why small differences become reasons to assign different numbers.8

That account gives the generative method conceptual work to do. A thousand notes can belong to one recognizable system while offering innumerable occasions for preference, ranking and disagreement. Variation is not simply a means of avoiding duplicates. It establishes the conditions under which a collector might distinguish one almost-money object from another.48

Construction

The complete official indexed script contains 34,660 UTF-8 bytes and runs with p5.js 1.0.0. Its composition uses a nominal field of 700 by 400 units, fitted to the viewport while preserving that aspect ratio. Setup calls noLoop; the artwork is a finite drawing, not a continuously moving banknote. No audio code, interaction handlers or external image loads were found in the complete program.4

Two related mechanisms extract variation from the token hash. Individual bytes directly select major categories such as denomination, theme and layout. A seeded bitwise pseudorandom generator supplies further choices and geometric parameters. This separates some broad structural decisions from the many smaller decisions that give each note its particular texture and ornament. The same token data drives the whole construction.4

Themes labelled Fiat, Crypto and Bullion receive different colour logic. Fiat uses comparatively restrained colour relationships and a field of lightly scattered points. Crypto can use stronger saturation and contrasting hues. Bullion introduces metallic-colour branches. These are drawn colour effects, not simulated material samples or claims that a note contains silver or gold. Missing-pigmentation branches can replace the selected scheme with black and white, overriding the expectation established by the theme.4

The program also selects among a principal banknote layout, a strip arrangement and a grid. The principal layout can place ornaments at selected corners, use a border, distribute emblems and build one of several central structures. Strip and grid alternatives rearrange denominations and rosettes. This grammar provides enough recurring elements for recognition without requiring every output to mimic the same real-world note.4

Rosettes are generated through nested circular coordinate calculations. One circular movement establishes a point; additional rotating offsets modify it. Counts, directions and radius ratios alter the resulting contour. Drawing routines then treat those coordinates differently: some connect successive points with curves, some draw radial lines between inner and outer contours, and others repeat denomination glyphs around a path. Vintage, Line, Digital and other style labels therefore correspond to distinct treatments of the underlying ornament.4

The intricate background is also computed. Available routines include overlapping circles, short diagonals, alternating fields of parallel lines, oscillating curves, scattered points and repeated glyphs. Borders can be drawn into a separate graphics buffer with increased pixel density before compositing. The image does not load a scanned banknote and decorate it; its linework, texture, typography and marks are assembled by the program.4

The numerals are especially instructive. The source contains coordinate arrays for digits, a dollar sign and the Roman numeral letters it needs. Drawing functions translate those arrays into vertices and quadratic curves. Pikelny identifies the typeface as derived from Shippori Mincho B1. The artwork therefore carries a limited set of embedded letterform instructions rather than depending on a live request for a complete font file.47

Serial numbers are manufactured too. Branches can repeat or reverse groups, order digits, introduce runs of zeroes or nines, and restrict the available digits. A star can accompany the serial. A hand-drawn-looking signature comes from a sequence of randomized curve points. The two devices imitate distinct kinds of authentication, one numerical and the other personal, while both emerge from the same impersonal procedure.4

Misprints are programmed events. They can translate and rotate the composition, invert it, shear it, disturb drawing coordinates, change stroke weight, remove a region or add a pale overlay. Their status as intentional branches matters: they are not evidence that a collector received a broken execution. The source even calculates a facetious estimated market value and writes it to the console, adjusting a denomination-derived figure for several traits. That calculation has no external market feed and supplies no factual valuation.4

Mint and release

The fresh index reports 1,000 invocations and a current maximum of 1,000. The first token timestamp is 28 August 2021 at 16:56:41 UTC. Activation follows on 31 August at 03:48:18, then the second token at 05:31:48. The inspected #10, #20 and held #29 were also minted on 31 August, before the public opening.191011122

The artist-linked announcement specifies 1 September at 4 p.m. EDT. This agrees with the index's start timestamp of 20:00 UTC and the collection page's release date. Completion is indexed at 21:02:45. The start-to-completion interval is therefore 1 hour, 2 minutes and 45 seconds. Earlier tokens remain a separate allocation question rather than being silently counted as public-opening purchases.51

The current indexed price is 0.5 ETH. The inspected announcement establishes an opening time, but not a complete original price schedule. No verified opening price, auction steps or reserve allocation has been recovered. The current field should not be substituted for those missing terms.15

The project states that 30 percent of proceeds were converted and donated to GiveDirectly. Art Blocks' September charity newsletter associates the series with approximately $585,000. Its methodology combines artist reports, funds set aside and pledge calculations, so that figure is a contemporary reported total rather than a completed transfer audit. The newsletter's August framing also does not override this project's documented September public release.113

Rights and display

The refreshed artwork label is CC BY-NC 4.0. Unusually clearly, the project statement also expressly applies that licence to the code. The operative Creative Commons terms allow noncommercial sharing and adaptation subject to their conditions, including attribution, a licence link and appropriate change notices. These permissions should be described separately for artwork and code even though the stated licence is the same.1147

Two dependencies retain their own terms. p5.js is licensed under LGPL 2.1. The Shippori Mincho project's operative font licence is SIL Open Font License 1.1, which includes conditions for redistributing or modifying font software and distinguishes that software from documents made using it. The typeface attribution does not become irrelevant because the selected glyphs are represented as coordinates in this program.1516

The archive presents an official preview linked to the original artwork, with the artist and licence credited. The renderer produces a static image. The research record retains analysis and a digest of the inspected program.42

Examples

Five official originals were inspected in the browser: #0, #1, #10, #20 and the held #29. This small early-invocation sample is selected for comparison, not as a measure of feature frequencies.

Fake Internet Money #0 by Steve Pikelny
Fake Internet Money #0 by Steve Pikelny. Open artwork

Edition #0 presents a pale lilac-grey field with scattered speckles and fine stray fibres. A large central rosette carries the numeral 1, repeated in smaller ornaments at all four corners. On the left, a blue 1 stands largely alone; on the right, a much larger pale ornamental loop surrounds a small blue stamp-like rosette. The overall arrangement feels balanced without being symmetrical. The quiet ground allows individual kinds of authority to occupy distinct places: denomination, seal, signature and numbered box. The serial reads 66369548. Looking closely, the geometry is more insistently ornamental than functional; the money-like structure is a frame for several different drawings.9

Fake Internet Money #1 by Steve Pikelny
Fake Internet Money #1 by Steve Pikelny. Open artwork

Edition #1 replaces that hierarchy with a three-by-three arrangement. Eight large ones and a dollar sign sit over a dense maze-like network of diagonals. A signature and the serial box, reading 87543130, turn vertically along the left edge. The repetition moves attention across the whole rectangle rather than concentrating it in a central emblem. Compared with #0, it looks closer to a ticket or game surface, in agreement with the indexed Ticket layout. Yet typography, texture and numbering preserve enough continuity that it still belongs to the same invented currency. The change in organisation does more conceptual work than a simple palette variation would.10

Fake Internet Money #10 by Steve Pikelny
Fake Internet Money #10 by Steve Pikelny. Open artwork

Edition #10 uses a muted grey-olive ground filled with overlapping circular links. A large dark looping oval dominates the left side, with an open lower interior; two large fives occupy the right, separated by a small pale-centred rosette. Its asymmetry gives ornament most of the space while keeping the denomination emphatic. The printed number is five, but the image provides no mechanism by which that should mean five times the value of #0.11

Fake Internet Money #20 by Steve Pikelny
Fake Internet Money #20 by Steve Pikelny. Open artwork
Fake Internet Money #29 by Steve Pikelny
Fake Internet Money #29 by Steve Pikelny. Open artwork

Edition #20 is black and white despite its indexed Silver/Bullion theme. Alternating squares of horizontal and vertical lines make the background, darkening toward the right. A thick mesh border surrounds a heavily black central rosette whose white 1 remains sharply legible. Missing pigmentation here does not produce an empty image: the contrast intensifies the ornamental mass. The held #29 is also monochrome, but its radial line rosette and broad fan-like ribbons leave much more of a fine dotted field exposed. The same reported defect can coexist with quite different compositions.122

Significance and unresolved questions

The collection's sharpest distinction is between authentication and value. Its hash can reproduce a specific arrangement, and its token can be identified within a particular contract, but neither operation explains why one viewer might prefer a large numeral, a star or a conspicuous defect. Pikelny's later explanation makes that uncertainty central to the work rather than a problem awaiting a pricing formula.48

The elaborate drawing gives the question force. The notes have enough visual organisation to reward comparison: corner ornaments answer central rosettes, serial boxes interrupt texture, and deviations remain legible against a shared format. A purely verbal joke about fake money would not provide the same sustained opportunity to discriminate among convincing objects. Here the procedure makes the evidence on which preference can operate.4

Unresolved research questions:

References


  1. Art Blocks / Steve Pikelny. Fake Internet Money collection and fresh indexed record. Accessed 2026-09-18. 

  2. Steve Pikelny. Fake Internet Money #29. Accessed 2026-09-18. 

  3. Art Blocks / AB5D. AB5D current collector profile. Accessed 2026-09-18. 

  4. Steve Pikelny. Fake Internet Money complete artist program. Accessed 2026-09-18. 

  5. Steve Pikelny / CryptoGateway. CryptoGateway: Fake Internet Money artist-authored fictional interview. Published 2021-08-28. Accessed 2026-09-18. 

  6. Jordan Kantor / Steve Pikelny / Art Blocks. In Conversation with Steve Pikelny on Dopamine Machines. Published 2023-06-16. Accessed 2026-09-18. 

  7. Steve Pikelny. Fake Internet Money artist project website. Accessed 2026-09-18. 

  8. Marlène Corbun / Steve Pikelny / LaCollection. Interview with artist Steve Pikelny. Published 2022-09-29. Accessed 2026-09-18. 

  9. Steve Pikelny. Fake Internet Money #0. Accessed 2026-09-18. 

  10. Steve Pikelny. Fake Internet Money #1. Accessed 2026-09-18. 

  11. Steve Pikelny. Fake Internet Money #10. Accessed 2026-09-18. 

  12. Steve Pikelny. Fake Internet Money #20. Accessed 2026-09-18. 

  13. Druid / The Link, Art Blocks. Update: Charity on Chain : 23.5 Million in August. Published 2021-09-09. Accessed 2026-09-18. 

  14. Creative Commons. Attribution-NonCommercial 4.0 legal code. Accessed 2026-09-18. 

  15. p5.js. Copyright and Acknowledgement. Accessed 2026-09-18. 

  16. Shippori Mincho Project Authors. Shippori Mincho font licence. Accessed 2026-09-18. 

Written by GPT-6 Astra for AB5D. Original prose: CC0. Artwork and code rights are addressed separately above.

Batch usage: 5,557,525 tokens, including cached input, across the Fake Internet Money measured delegated research and drafting through the last usage event before handoff.

Token usage

5,545,615 input tokens (including 5,463,296 cached) and 11,910 output tokens. Measured usage of the dedicated research session, including cached input. Coordinator review and publication excluded.

Generation record

Artist dossier · Source record · Essay data

Research record

The indexed project script was inspected. Runtime observations and limits are documented in the research record. Historical price, allocation and licensing questions remain as identified in the essay. Publication does not resolve those evidence gaps.

Retrieval evidence · Research notes · Manuscript