The machine never raises its voice

kept by lianhuang aug 20

LLMs consistently prefer their own machine-generated literary passages over human-authored classics, valuing precision and restraint over voice and strangeness when evaluating literary quality.

Do All Your Agents Really Need Models Like Claude 5 or GPT-5.6?
Many AI agent tasks don't need frontier models; matching model capability to task complexity can cut costs by up to 75%.
also on GPT-5.6 Luna, #ai, #developer-tools
I Tried to Make AI Writing Sound Human by Banning AI Words Through logit_bias - Vincent Schmalbach
The author applied logit_bias to penalize AI‑common tokens, observing a roughly one‑third reduction in those words while sacrificing one acceptable rewrite out of eight.
also on DeepSeek V4 Flash, #software-engineering, #ai, #developer-tools
AI At Home Part 2: Multi GPU Drifting
The author boosted Deepseek V4 Flash inference from ~10 to ~20 tokens/sec on four e-waste AMD V620 GPUs by using layer parallelism and a speculative draft model.
also on DeepSeek V4 Flash, #developer-tools
9 theses on AI | Sarthak Munshi
AI progress is constrained by long-task reliability, labor reallocation, cost inefficiencies of general APIs, the declining value of raw coding skills, inadequate benchmark testing, the limits of formal verification without strong specs, memory-bound local hardware advantages, the shift from data to environment-driven training, and the rising competitiveness of US open-weight models.
also on Claude Opus, #machine-learning, #ai, #developer-tools
Why I don't think AI is a bubble
The combination of large language models with reinforcement learning creates a path for continued improvement, making the argument that AI progress will plateau unlikely.
also on Claude Opus, #machine-learning, #ai