Hacker News with Generative AI: Text Generation

Strengths and limitations of diffusion language models (seangoedecke.com)
Google recently released Gemini Diffusion, which is impressing everyone with its speed. Supposedly they even had to slow down the demo so people could see what was happening. What’s special about diffusion models that makes text generation so much faster? Should every text model be a diffusion model, going forward?
Refactoring Clojure (orsolabs.com)
This article is based on Writing Friendlier Clojure by Adam Bard, where he shows his approach at refactoring some Clojure code that implements an order-1 word-level Markov text generator.
Qwen2.5-Omni Technical Report (huggingface.co)
In this report, we present Qwen2.5-Omni, an end-to-end multimodal model designed to perceive diverse modalities, including text, images, audio, and video, while simultaneously generating text and natural speech responses in a streaming manner.
A new semantic chunking approach for RAG (gpt3experiments.substack.com)
As we saw in my last blog post, there is a shape for stories.
Llms.txt (llmstxt.org)
LongWriter: Unleashing 10k Word Generation from Long Context LLMs (arxiv.org)
Claude 3.5 Sonnet Reproduces BIG-Bench Canary String (lesswrong.com)
Can language models serve as text-based world simulators? (arxiv.org)
Show HN: Technical diagrams from text: HTTPS://text2diagram.com/ (text2diagram.com)