We’re very accustomed to portraits of people—we see them all the time. But can portraiture be extended to digital entities such as files, folders, programs, and information systems?

A good portrait, it is commonly believed, captures the inner world of its subject.

Enter The Machine is a collection of algorithmically and human generated (“algo-human”) portraits that capture the complex inner world of digital entities.

This series of work asserts that not only can portraiture be expanded to include digital entities, but doing so will give us new ways of seeing and understanding the digital world around us.

Harkening back to the centuries-old tradition of subjects coming to an artist’s studio to have their portrait made, a portrait studio for digital entities is a contemporary version of this model–except the portraits produced are of digital files, folders, programs, and information systems rather than people.

Similar to a traditional portrait studio, Enter The Machine 3.0 creates portraits of two kinds of subjects:

1. Culturally and historically significant: Portraits of Famous Files
2. Personally significant: Portraits of People’s Files

File portraits are created via a series of steps:

• Transforming the file(s) into binary code (0s and 1s)
• Run the binary code through custom-built software that interprets the binary data as colored pixels on a geometric grid
• Manually transform the algorithmically generated results by shifting colors, modifying composition, adding a background, custom sound, and/or animation
• Deploy the visual portrait on the web

As noted above, the form of the portrait (shape and pixel distribution) is algorithmically generated while the coloring, composition, background, animation, and sound are created manually. The portrait creation process is “algo-human”–half made by man, half made by machine.

Portraits of Famous Files
We’re all familiar with portraits of famous people—we see them all the time. They’re typically those who have had a significant cultural or historical impact on society. Similarly, the collection Portraits of Famous Files, are portraits of digital files that have had significant historical or cultural impact dating from the start of the digital age in the 1960s through the present day. For example:

• All files in Edward Snowden’s NSA Data Leak (2007)
• SETUP.EXE for AOL 3.0 (1995)
• Zelda for original Nintendo (1986)
• An Information Management Proposal, by Tim Berner’s Lee (1983)
• The code NASA wrote for the Apollo 11 mission (1968–1969)

Feel free to make a suggestion for the inclusion of a famous or view the full collection here.

Portraits of People’s Files
This series consists of custom file portraits made from an individual’s or a group’s personal files. They are not famous, culturally speaking, but they tend to hold personal significance for the group or individual. It is the digital equivalent of going to a portrait studio with those close to you, except instead of bringing people, you’d be bringing digital files that are meaningful to you. Learn more about commissioning a file portrait.

At this time, ownership is only available for commissioned portraits, however this may change in the future. Sign up here for updates.

File portraits range from $2,500–$7,500 depending on demand, complexity, file size, number of files, and whether the portrait is animated and/or has custom sound design (like so).

You can get a rough idea of where in the above range your request would fall here. To get an exact quote, fill out the commission request form. There are two ways to pay.

1. Ethereum (preferred)
Paying via Ethereum is preferred because:

• Since one of the main deliverables is an NFT on Ethereum, paying with the same wallet that will receive the artwork is the most straightforward and secure method
• Provides transparency around pricing for prospective and future commissioners
• Avoids credit card fees

2. Credit card/PayPal/monthly installments
You can also pay via traditional payment methods, such as credit card, PayPal, or monthly installments, via the artist’s online shop (a custom order will be created). Notes:

• You will also need an Ethereum wallet address to receive the final work
• These methods will incur additional fees

The commission fee is just one of many factors the artist considers when accepting a commission, and it is by no means the deciding factor. As a general rule, the artist prefers to take on interesting commissions and does not automatically accept commissions offering higher fees. That said, there are economic realities the artist must consider, and as such, higher commission fees can sometimes help increase the chances of having one’s commission accepted.

The first thing to note is the minimum commission fee. After the minimum fee is met, commission fees should generally scale in relation to the complexity of the request:

• $: static portrait, consisting of a single file, that is not animated and does not contain sound, like this or this
• $$: animated portrait, consisting of a single or a few files, that does not contain sound, like so
• $$$: animated portrait, consisting of a single or a few files, that does contain sound, like this or this
• $$$$: animated portrait consisting of many files that contains sound, like so

Naturally, each individual will have to decide what to offer based on what they want and their financial situation. Please bear in mind, there are no returns.

A word of advice from the artist: it’s quite possible you may not like how the work turns out, so for everyone’s sake, don’t offer more than you can easily afford to part with. This is nothing if not an experiment, and not all experiments succeed.

When you commission an Enter The Machine work, you will be in direct contact with the artist’s studio and will get the following:

• Face time with the artist to discuss the commission and how to realize it
• Web-based file portrait, like this or this, deliverable as an NFT on Ethereum
• 24" x 36" signed print of the work (other sizes available upon request)
• Official documentation of the work–like this for images or this for videos (file will download automatically)–which is permanently stored on the Arweave blockchain
• Multiple images or videos for social media posts
• Have questions about how to display the artwork at home? See here.

What you do not get

One is, of course, free to choose any file(s) one likes, but generally speaking, the artist recommends selecting files that are personally significant or meaningful. A good rule of thumb is whether you could talk about why a file or files are important to you for at least one minute. Some examples may be a script or important paper you wrote, a book that never got published (or did!), a program you coded, a video you made, a site you built, a voice message you received, a photo of significance, etc.

A portrait can be created from literally any digital file or files. Currently, the only limitation is that the size of any individual file not exceed 20MB. Note: compositions of multiple files, where each file is less than 20 megabytes, do not pose a problem. Work is currently being done to increase the individual file maximum size.

Start the process by reviewing the commission page. There’s a form to complete at the bottom.

The creation process is a combination of algorithmic and manual work, all of which is done by the artist in his studio in Brooklyn, NY.

The process varies based on the portrait but, in broad strokes, it consists of steps below which generally take between 4–12 weeks. Here’s how it typically goes:

1. Decide. The process begins by deciding on the digital entity that you would like a portrait of (file, folder, program, nft, system, etc.) See “What kind of file should I choose?”
2. Request. Fill out the commission request form. Be sure to provide screenshots and discuss the significance of the file. The more details, the better.
3. Wait. Most commission requests, unfortunately, cannot be accommodated due to the labor-intensive nature of the creation process and the limited bandwidth of the artist. See “Will my commission request be accepted?”
4. Pay. If your commission request is accepted, the artist’s studio will reach out via the contact info you supplied, and you will be asked to pay the commission fee upfront. Please note the fee is not refundable. See “How much does it cost and how can I pay?”
5. Meet. Schedule a 15-30 minute meeting over Zoom to discuss the commission with the artist. This is an opportunity to get to know the artist and talk about the personal significance of the file(s), as well as the criteria by which the portrait will be successful.
6. Wait. Work will begin once the artist’s studio has all the files and information it needs. This process can take anywhere between 4–12 weeks.
7. Receive. As the commissioner of the work, you will receive:
   • File portrait as an NFT on Ethereum (this requires an Ethereum wallet address)
   • Signed 24" x 36" print
   • Link to download PNG or MOV documentation
   • Images or video for social media posts
   • Link to the portrait in the Portraits of People’s Files Collection on v3.enterthemachine.io

The time it takes to complete a file portrait depends on the complexity of the request, whether it is animated, has custom sound, entails lots of files or large files, and current demand.

Generally speaking, portraits are completed within 4–12 weeks.

There are currently two types of Enter The Machine file portraits: individual and group.

• Individual portraits: consist of only a single file. Examples:
  • SETUP32.EXE
  • bitcoin.pdf
  • the-declaration-of-independence-of-the-united-states-of-america-by-thomas-jefferson.txt
• Group portraits consist of multiple files. Examples:
  • snowden-archive: 6 directories, 474 files
  • Netscape Navigator 9.0.0.6.exe: 35 directories, 186 files
  • proposal.html: 1 directory, 5 files

To create a group portrait of a folder of files or an application, simply zip and submit the folder.

Yes. It’s a quick one-on-one to help the artist get a sense of you, the relationship to your file, and your expectations for the portrait. Good portraits are built on good foundations, which only come with actually knowing the subject, hence a meeting.

It’s been less than 6 weeks since I submitted…
It’s understandable that, nowadays, waiting 6 weeks to get a response from an internet-based project seems overly long. Bear in mind that this is a small-time operation led by an independent artist and not a large. Please continue waiting. Your patience is very much appreciated during this time.

It’s been more than 6 weeks…
If you still haven’t heard anything after six weeks, the recommended course of action is to submit a new commission request where either the files, descriptions, or commission fee is different from the original request.

There are two general reasons why a portrait request may not be accepted and several case-specific reasons.

General reasons
Firstly, it’s worth bearing in mind that most file portraits are created by just one person (the artist, Eric Corriel) so resources are highly limited and the majority of requests simply cannot be accommodated due to the nature of the process, which is not automated and requires a lot of manual work and individual attention.

The most likely reason a commission is not accepted has nothing to do with the commission; it’s just that the artist is fully booked.

Secondly, when deciding which commissions to accept, the artist is thinking about how the work resulting from the commission would fit into the existing collection and whether it offers the artist—and viewers—new perspectives compared to what has been done before.

Case-specific reasons
The first things the artist looks at when considering a commission are the provided screenshot images and the personal story that explains why the commission is being requested in the first place. The artist is looking for a combination of compelling files, screenshots, personal stories, ideas, etc., as well as diversity in terms of file types and geographic location of the commissioners.

The next factor that can play a role is total file size. Generally speaking, the larger the total file size of the commission, the more resource-intensive and time-consuming the creation process will be. As a general rule, the artist is more selective of commissions totaling more than 50 megabytes of data.

Lastly is the commission fee. The fee is just one of many factors the artist considers when deciding to accept a commission, and it is by no means the deciding factor. As a general rule, the artist prefers to take on interesting commissions and does not automatically accept commissions offering higher fees. That said, there are economic realities the artist must consider, and as such, higher commission fees can sometimes help increase the chances of having one’s commission request accepted.

What to do next
If you’re still interested in an Enter The Machine file portrait, the best thing to do is to review the content above, rethink your submission, and submit a new request.

In centuries past, portrait painters would, on occasion, revise a portrait upon the patron’s request. Enter The Machine portraits follow this tradition, and revision requests may be honored, though this would be solely at the artist’s discretion.

Unfortunately not. The buyer must acknowledge they are taking a risk when they commission work from an artist who has his own vision for how things should turn out. Satisfaction is not guaranteed.

Sure thing—you do you!

Enter The Machine file portraits, like so much digital art, are created at the outer edges of what the technology of the day can provide. To prepare for the eventuality that current technical standards may not be operable in the future, the documentation of each Enter The Machine work is preserved on the Arweave file storage blockchain, which is designed to store information for the next 200 years.

To be shown in their internet-native format, Enter The Machine works requires a host of technologies working together:

“Client-side” technology requirements (i.e., what a viewer needs to see the work):

• Broadband internet
• Web browser that supports JavaScript ECMAScript 2023, CSS 3, and HTML5

“Server-side” technology (what’s needed to serve the work)

• Enterprise-grade, object-based storage capable of storing hundreds of millions of files, which in this case is Amazon’s S3 storage service
• Enterprise-grade content delivery network (CDN) to serve millions of files efficiently across the world, which in this case is Amazon’s Cloudfront CDN service
• Storage platform to host necessary HTML, JavaScript, and CSS files that glue the appropriate tiles together to form the portrait at a particular zoom level, which in this case is the decentralized Arweave file storage blockchain

While the artist’s studio will do what it can to mitigate future incompatibilities, the buyer must nonetheless accept some risk when owning technology-based art with known external dependencies that it may not display as intended in the future—conditions that widely apply to most forms of digital art.

If you’re wondering why these works cannot be fully stored onchain, see here.

Enter The Machine 3.0 works support varying levels of interactivity depending on their technical foundation. The best way to see what kind of interactivity a particular work supports is to consult its Interactivity section. Below is the current list of possible interactions.

Play/pause audio (for those that have audio)

• Desktop: sound icon if viewing on v3.enterthemachine.io or s key
• Touch devices: tap and hold for two seconds then play
• Note: sound is muted by default by all web browsers and can only be unmuted by initiating the actions above

Stop/restart animation (for those that are animated)

• Desktop: spacebar
• Touch devices: tap and hold for two seconds then play

Zoom

• Desktop: use a mouse/touchpad that supports or simulates mouse-wheel functionality
• Touch devices: pinch and zoom
• Note: if the portrait is animated, it needs to be stopped before it can be zoomed

Pan

• Desktop: click, hold, and drag
• Touch devices: tap, hold, and drag
• Note: if the portrait is animated, it needs to be stopped before it can be panned

Rotate

• Desktop: Opt (Mac)/Alt (Windows) + click, hold, and drag
• Touch devices: pinch and rotate
• Note: if the portrait is animated, it needs to be stopped before it can be rotated

Fullscreen

• Desktop: f key (note: does not work on all 3rd-party platforms, such as OpenSea)
• Touch devices: not available

Despite appearances, Enter The Machine works are neither strictly images nor videos. Rather, they are complex webpages consisting of HTML, CSS, JavaScript, and hundreds of thousands of small image files, all working together to produce what looks like a single image or animation.

Enter The Machine file portraits are built on technology similar to Google Maps. Each portrait consists of hundreds of thousands, and sometimes millions, of 256-pixel x 256-pixel tiles that are swapped in and out depending on which section of the portrait is being viewed and at what zoom level.

This allows Enter The Machine works to be:

• Responsive: display well across all devices, from mobile to billboard-size screens, in any orientation (portrait, square, or landscape).
• Interactive: user can zoom, pan, and rotate.
• Dynamic: works can be programmed to incorporate randomness and respond to user-driven events.
• Scalable: the portraits themselves can be of any size, from 100 pixels wide to hundreds of thousands.
• High resolution: portraits can be zoomed in without sacrificing integrity.
• Support future display devices: portraits are created at very high resolutions to help them adapt to future, higher-resolution devices. Additional layers of resolution can be added to support future devices.
• Lightweight: progressive loading (i.e., loads in piecemeal fashion) to accommodate varying bandwidth and Internet congestion conditions. As such, a naturally occurring element of visual randomness gets introduced as the loading pattern is slightly different each time.
• Accessible: all Enter The Machine works contain visual and aural textual descriptions accessible to screen readers and other assistive devices.

It is well known among conservators of digital art that preserving complex digital works in their original form over significant periods of time is extremely difficult, and sometimes impossible.

Digital artists make work using cutting-edge hardware and software that, several years hence, may very well be relegated to the technological trash bin of history, which would render the work unviewable in its native format should the original components stop functioning.

Whether this will happen to Enter The Machine works 5, 10, 20, or 30 years hence is anyone’s guess. Still, to mitigate against this possibility, documentation (in the form of standard image and video files) for each work is stored on the Arweave file storage blockchain, which is designed to preserve files for the next 200 years. Image example here, video example here (file will download automatically).

Documentation for each work can be found under its “Digital Preservation” section. In the event an internet-native work no longer functions, the onchain documentation can be used in its place.

Scale

Number of digital entities, typically meaning files or folders, in a portrait

Diversity

Logarithmic scale, from 1-10, of how many different file types are in the portrait.

Diversity is computed by the following forumla:
(1 + 9 × (log10(num_file_types_in_portrait) / log10(100)))

Dominance

The most prominent type of file in the portrait (for example, .jpg, .doc, .html).

Epoch

Decade in which the file was created.

Density

Total number of bytes represented in the portrait.

Sound

Whether the portrait includes sound.

Time-based

Whether the portrait is static or changes over time.

Portrait Type

Whether the portrait is of an individual digital entity or a group.

While all code (HTML, CSS, JavaScript) and supporting files (mp3, mp4, etc.) are stored onchain (on the Arweave file storage blockchain), the millions of 256x256 image tiles that comprise each file portrait are stored on Amazon’s centralized storage solution, S3.

The two main bottlenecks to storing millions of files onchain are upload times and indexing.

The tooling required to upload millions of files to existing cloud storage infrastructure, such as Amazon’s S3 service, is quite advanced and supports multi-threaded, parallelized uploading. Tools like s5cmd can upload millions of files in hours. The tooling to upload to decentralized storage is primitive by comparison. Currently, most decentralized storage services can only handle serialized uploading (one file at a time), which would require days or weeks for Enter The Machine works, which is not practical at present.

Regarding indexing, which refers to quickly finding and serving files, most decentralized storage platforms are nowhere near as efficient at serving folders containing millions of files as centralized alternatives that have had decades to scale up to such demands and can now do so effortlessly.

Yes. The algorithm that creates the file portrait is deterministic, meaning given the same input (data contents of the file) it will produce the same output. However, if any of the data changes, even just slightly (imagine two Word documents that contain identical text except for one word), the resulting portraits will be visually different, albeit only slightly in this case.

Never say never, but no, I very strongly believe it is not possible to reconstruct the original file from the visual portrait for the following reasons:

• Data is not perfectly preserved in visual form due to the fact that pixels representing some data overlap other pixels representing other data
• There aren’t definitive visual markers denoting the first and last pieces of data, which means many, many possible combination of start, end, and middle points would need to be tried to reconstruct the data in the correct order. While this may be feasible for small portraits, it would not feasible be for large portraits.
• The most significant reason: There are dozens of intermediary, manually-derived “color correction” layers between the algorithmically generated result and the final work, which effectively severs the direct connection between the data and its visual representation. There is a human muddling around in the loop, so to speak, which makes the process virtually impossible to reverse-engineer.

Enter The Machine works are meant to be displayed on digital screens in portrait orientation. There are two main ways of getting the work onto a digital screen.

1. Upload static assets to a digital frame (PNG or MOV files)
Probably the easiest way to go, provided one has a digital screen meant for displaying artwork (such as a Meural), is to load the .mov or .png documentation file onto the device. These files are available to download via the “Digital Preservation” section of any work, like so.

Pros: Easy to implement

Cons:

• Depending on the screen’s resolution, the image may not be as sharp as when it is viewed natively via a web browser. Pixelation would be more pronounced with videos than with images.
• If the screen’s aspect ratio and orientation do not match the image’s—9:16 (portrait orientation)—the work will likely either be letterboxed or cropped.

2. Native web format
If one has a digital screen with a web browser that can display URLs, then one could show Enter The Machine 3.0 works by pointing their device to the following URL: https://v3.enterthemachine.io/collections/portraits-of-famous-files/embed/1

Notes

• Replace embed/1 in the URL above with embed/{portrait_id}, which can be found in the URL of the portrait. For example, the {portrait_id} of https://v3.enter-the-machine.io/collections/portraits-of-famous-files/019-ransomware-wannacry-2-0 is 19.
• Depending on the collection, you may also need to replace portraits-of-famous-files with portraits-of-peoples-files.

Pros: Best image quality, aspect-ratio agnostic

Cons: May not be supported by non-industry standard browsers that may be found on TVs or other display devices (anything that isn’t the latest versions of Chrome, Firefox, or Edge)

3. Prints
See the following section

The studio is currently exploring different types of print solutions. If you are interested in a print, please fill out this form.

• File portraits are slow to render and animate on Apple Vision Pro, Vision OS 2.0