diff --git a/content/authors/valknar/index.md b/content/authors/valknar/index.md
index f908f02..ec74561 100644
--- a/content/authors/valknar/index.md
+++ b/content/authors/valknar/index.md
@@ -1,14 +1,14 @@
 ---
 title: "Valknar"
 name:   "Valknar"
-bio:    "Valknar works with diffusion models as a compositional tool, creating dark surrealist visions that challenge the language of machine vision and disrupt comfortable readings of the generated face."
+bio:    "Valknar treats diffusion models as a engine of contradiction — pitting the cute against the grotesque, the opulent against the decayed, and the familiar against the deeply wrong. The results are cinematic concept images that feel like stills from films that should not exist."
 social:
   instagram: "valknarix"
   x:         "valknar100"
 ---
 
-Valknar approaches AI generation as a negotiation — between the model's statistical memory of human aesthetics and a deliberate effort to pull those aesthetics toward the unfamiliar. The results are portraits and figures that feel simultaneously recognisable and wrong.
+Valknar approaches AI generation as a collision course. The work takes familiar cultural touchstones — video game icons, children's characters, Renaissance still life — and forces them through genre frameworks they were never designed to survive: survival horror, Gigeresque biomechanics, neo-Tokyo neon, Burtonesque gothic.
 
-Working primarily with portrait and figure generation, Sable uses seed manipulation, latent space traversal, and post-processing to extract images that resist the smoothness the models default to.
+The technique is less about coaxing beauty from the model and more about engineering productive discomfort. Vivid colour is deployed against desolation. Opulence is buried in decay. The cinematic framing is deliberate — wide shots, hero angles, macro details — borrowed from the language of film concept art and applied to subjects the camera would not normally linger on.
 
 Based in Stuttgart.