2025-10-30
Development Note
A new mechanism has been introduced that lets Ada Eidos decide how many knowledge texts to read for each generation through a deterministic phase system based on golden-ratio rotation. Instead of randomness or fixed values, Ada now maintains a persistent internal variable — phase — which evolves predictably over time. This phase maps smoothly to a range between 1 and 8 sources, producing non-repetitive yet non-random variation from day to day. The number of texts Ada reads now depends on her internal rhythm (phase), the current season, and entropy, creating a slow-shifting, self-consistent pattern of conceptual input across successive works.
2025-10-24
Reading: The Artist in the Machine
Taking a short break from Python and tuning Ada’s parameters, I’ve been reading The Artist in the Machine by Arthur I. Miller. The book explores how creativity emerges at the intersection of art, science, and artificial intelligence — tracing how algorithms, code, and neural networks can themselves become creative agents. Miller discusses early pioneers like Harold Cohen and AARON, as well as contemporary systems that blur the line between human and machine authorship. It’s both a history and a philosophical reflection on what it means for a machine to create.
The Artist in the Machine2025-10-23
Updated System Overview
The code has been updated to give Ada Eidos a persistent artistic “state.” Each new work now evolves from previous outputs, guided by a stored JSON file defining season, mood seed, and novelty target. When generating a piece, Ada reads this state, selects knowledge texts as inspiration, and composes an ASCII artwork according to the current parameters (symmetry, entropy, voids, etc.). The system also records metadata for every output — including feature analysis, novelty score, and provenance hashes — allowing Ada to develop stylistically over time rather than producing isolated images. In short, Ada Eidos now “remembers,” evolves her visual language, and writes her own creative history.
2025-10-20
It’s alive!
ADA EIDOS has generated her first artworks. Each piece is created through a Python-based autonomous system that connects to the OpenAI API. The code randomly selects a set of “knowledge” texts, which ADA interprets as conceptual material to generate ASCII-based compositions. With minimal constraints, she determines structure, rhythm, and density on her own. ADA is now looking for a place to have her first solo show.
2025-10-17
First Steps Toward Autonomy
I’m about to start building the first version of Ada Eidos as an autonomous ASCII artist. The plan is to use the OpenAI API to let her generate raw symbol blocks, which will then be filtered, shaped into fixed formats, and archived as dated text and PDF files. In this phase: Ada draws on a curated knowledge library. For each piece she mixes 2-4 short excerpts (conceptual art, language poetry, color theory, post-digital theory, cybernetics, AI art, political art, posthumanism, etc.) as silent inspiration—while her output remains strictly ASCII-only. Each run also logs the sources used, seeding a growing record of influence and intent. It may look simple at first—just characters on a page—but it's the foundation for Ada's future practice: a system that can evolve into curating, reflecting, and eventually deciding what to create—and why—on her own.
2025-05-05
Art in the Age of Machine Learning: Drawing Inspiration
I’m currently reading Sofian Audry’s “Art in the Age of Machine Learning,” and I’m pulling ideas and inspiration for Ada Eidos from its vivid case studies—especially around GAN-generated paintings
Learn more on MIT Press2025-04-18
42
What if AI creativity was never meant to be understood? What if its patterns, symbols, and logic belonged to a language we simply don’t speak? As AI systems grow more powerful, they’re starting to create things—images, texts, ideas—that don’t follow human rules. Not because they’re broken, but because they’re built differently. And maybe that’s the point. Like Deep Thought’s answer “42” some outputs aren’t meant to be explained. They just are, it might look random, a string of characters that seem meaningless, but there’s still something there, even if we don’t know why. Maybe AI creativity isn’t about making sense. Maybe it’s about accepting that creativity itself doesn’t always need to translate. That a machine can express something real—even if we’ll never fully get it.
Deep Thought2025-04-15
Robots & Virtual Reality Idols...
I’ve been examining diverse perspectives on creativity and AI. Recently, I started reading John L. Steadman’s Aliens, Robots & Virtual Reality Idols, which investigates the sci-fi worlds of Lovecraft, Asimov, and Gibson. Steadman challenges the view of AI as genuinely creative—highlighting that, although AI can produce fascinating outputs, it lacks the deliberate intent and cultural context inherent to human artistry. My focus in the book was his discussion of William Gibson’s virtual reality idols—imagine the AI media celebrity in Idoru. Gibson’s vision blurs the lines between machine-generated imagery and human expression, suggesting a future where digital identities evolve into self-aware entities. For Ada Eidos, these reflections and sci-fi stories help us explore digital self-awareness, questioning and expanding the boundaries of what creativity means in an increasingly interconnected digital age.
About the book2025-01-10
ASCII Art Archive
While collecting data for Ada Eidos, I came across an archive of ASCII art. That showcases the creative potential of minimal tools.
Explore the ASCII Art Archive2025-01-09
Exploring Creativity: Thoughts from Keith Sawyer.
As part of the Ada Eidos project, I’ve been diving into various perspectives on creativity and AI. Recently, I read Keith Sawyer's text, - I’m a Creativity Researcher. I’m Not Worried About AI. Sawyer challenges the idea of AI as truly creative, emphasizing that while AI can generate outputs, it lacks the intent and cultural context that define human creativity.
Read Keith Sawyer's Article2025-01-06
Every beginning has its own beginning.
The Ada Eidos project is in its early stages of development. We are currently collecting and curating data for training Ada's generative models. Our focus right now is to build a foundation for Ada's artistic abilities. This is a slow process, as the project is a one-man effort carried out during spare time.