TypeSafe's viral Jev model accused of uncited similarity to year-old Laya project
ECrispy · reddit · 2026-09-20
A developer questions whether TypeSafe AI's new classifier model Jev borrows from prior community work: an AI dev previously posted having built the same architecture a year ago as "Laya." The poster argues the similarities go far beyond surface level and that TypeSafe didn't cite or credit the original paper, inviting ML experts to weigh in.
More from Models
- Nat Lambert: OpenRouter data is more representative but still undersells closed model growth — natolambert · 2026-09-20
- OpenAI runs 10,000 agents to probe Navier-Stokes; DeepSeek V4.1 Flash debuts new architecture — DeepLearningAI · 2026-09-20
- Demo shows 'smart copy/paste' built on TypeSafe AI's Jev — menhguin · 2026-09-20
- Qwen3 27B quantization tests on 16-20GB VRAM: GSQ-RCO wins overall — Mordimer86 · 2026-09-20
- Dethrone: an open-source card game testing Jev, a model that outputs decisions instead of text — aaddrick · 2026-09-20
- Step-5-Preview BF16 weights leaked early, community mirrors fork on Hugging Face — External_Mood4719 · 2026-09-20