grebenshi — work index

grebenshi — work index

  • inpainting
  • obsolete
  • prints
  • gazecom

    gazeCOM is a complete rewrite — in React and TypeScript — and a substantial expansion of gengaze (2025), the local, browser-based environment for saliency-guided image generation, introducing a new VLM mode, an extensive LLM system overhaul and weighted pools for workflows and prompts.

    The original modular, self-contained tool for experimentation and hybrid control has grown from an interaction-focused paradigm into a fully-fledged "supergenerative" application, introducing a different approach to image processing. Alongside the classic interactive and algorithmic inputs — gaze, hand and cursor tracking, computer-vision prediction, and automated roamers — it now adds agentic, self-referential vision guidance.

    The premise is unchanged: a stream of coordinates is gathered into a heatmap, reduced in turn to its center of mass, and fed into the generative workflow together with the image to be processed. The COM steers the region that gets cropped and, in composite mode, the direction in which the canvas drifts and expands (making gazeCOM, in effect, an automated outpainting machine).

    With the introduction of the new VLM mode, at the close of a generation the returned frame can be sent back to a local vision model, whose reply — a single coordinate, drawn from an editable instruction — becomes the input point for the next cycle. The instruction is written in plain language: the most salient subject, an empty area, a weakly resolved region (and could itself be agent-driven). As the image changes, the next reading moves with it, so the system traces a self-rebalancing path through its own output rather than being influenced from the outside.

    Where gengaze opened an exploration of pre-cognitive space through gaze tracking in an enquiry about the limits of automation in the creative process, gazeCOM turns its focus onto machinic vision itself, becoming, in a sense, autopoietic.

    With language and vision both delegated to Ollama, image generation to ComfyUI across modes, gazeCOM becomes, in effect, a substantially multi-agent-driven pipeline.

    Now publicly available and open-source.

    Find it here.

    Watch on Vimeo (the player requires JavaScript).

  • neu/n
  • pyfa

    Pick Your Favorite Artist is a project hosted on EmProps, a platform for long-form generative AI. It's a collection of 105 outputs from a workflow that incorporates a second generative layer — a p5js-based image-to-image algorithm — which effectively makes it supergenerative, as described by prof. Ira Greenberg: a rule-based system operating on top of other rule-based systems.

    The piece pays homage to nine painters: Hernan Bas, Cecily Brown, Dana Schutz, Nicole Eisenman, Peter Doig, Francis Bacon, Neo Rauch, Odilon Redon and Justin Mortimer. Their distinctive styles are embedded into prompts run through an SDXL pipeline, and the subjects strive for contrast, playing with the tension between those styles and a projected, ironic character reminiscent of hip-hop and gangster imagery.

    With 45 different "features" — parameters influencing the prompts through either words or images — spread across six variables, the potential output combinations amount to 47,385.

    The project was followed by the release of PYFA SHORT — a mini collection of 31 curated outputs.

  • graffuturism

    graffuturism is a lab for generative ai tooling, run from 2022 to 2024 on the models of the time — Disco Diffusion, Stable Diffusion, Midjourney, Flux.1 — before editing models existed. it used Dreambooth and LoRA training on datasets built from my abstract 3D work, whackom above all.

    its output was released through FormFunction and ExchangeArt, and now lives as a mixed collection on Mallow

  • ukiyo-ai

    ukiyo-ai is a blind drop released on Foundation in 2023. it originated as a study series using early generative ai models. find it on networked / raster

  • muñañyo
  • whackom
  • oculus
  • edit
  • painting
  • journal
  • bluescreen
  • mallow
  • objkt
  • superrare
  • vvv
  • raster
  • pnd