← Dashboard  /  News

2026-09-07

AI/Tech Digest 2026-09-07

AI & local LLM

1. 8 uncensored Qwen 3.8 27B variants, one base, 167 GPU hours - Abliterlitics [LAB]
   Provides comparative performance data on uncensored fine-tunes of a mid-size model relevant to local inference.
   https://www.reddit.com/r/LocalLLaMA/comments/1w8vx6w/8_uncensored_qwen_38_27b_variants_one_base_167/ (reddit.com)
2. Research acceleration: The view inside OpenAI [LAB]
   Discusses the use of coding agents to accelerate AI research, touching on agent frameworks and task complexity.
   https://openai.com/index/research-acceleration-view-inside-openai (openai.com)