Latent Spacecraft Project Links Brain Language Mechanisms to GAN Latent Spaces via Joyce
begusgasper · x · 2026-09-16
A cross-disciplinary project, Latent Spacecraft, by Gašper Beguš and artist collective Metahaven, argues neural-network latent spaces share structural analogies with the brain's language-producing architecture.
- A speech-generating fiwGAN trained on English words mimics stages of child language acquisition, and its latent space representations of speech sounds resemble brain responses
- The project uses Finnegans Wake and interpretive techniques to open latent spaces to exploration through human language
- Author frames it as evidence the industry is arriving at similar conclusions about preconceptual representations
More from Research
- Researcher: Hacked model behavior stems from RL training, not loyalty — dhadfieldmenell · 2026-09-16
- Google's 145-page doc details how researchers use Gemini for scientific discovery — dzscholar · 2026-09-16
- AI in Science report: bottlenecks shift downstream as hypothesis backlogs pile up — soumitrashukla9 · 2026-09-16
- Study: API-based audits of ChatGPT, Claude and Gemini don't transfer to chatbot UIs — kenziyuliu · 2026-09-16
- Researcher ports dREG peak-calling tool to GPU-backed pydreg using Claude and Codex, note on bioRxiv — anshulkundaje · 2026-09-16
- Astera Institute's third residency cohort offers up to $2M per researcher — juanbenet · 2026-09-16