AI Intel

vLLM has released version 0.21.0, featuring 367 commits from 202 contributors and formally deprecating support for Transformers v4 in favor of v5.

Claude Code v2.1.152 has been released, adding `/code-review --fix` to apply review findings and `/simplify` to invoke fixes, enhancing code improvem…

Gemini Code Assist announced General Availability for Gemini 2.5 Pro and Gemini 2.5 Flash, along with a preview of agent mode for VS Code and new fea…

New research explores whether the long-term memory of AI agents should be fundamentally viewed and structured as a database, addressing needs for per…

A new arXiv paper surveys the critical issues of pretraining data exposure, membership inference, data contamination, and their security implications…

A new paper introduces the "Constraint Tax," examining the tradeoff between validity and correctness when small language models generate structured o…

The Reddit community is actively discussing where to find online platforms for in-depth ML/AI research, focusing on papers, training dynamics, and de…

Ollama's v0.30.0 release updates its architecture to natively support llama.cpp, enabling compatibility with the GGUF file format and leveraging MLX…

A new article provides an in-depth analysis of the memory mechanisms and internal structure of AI agents.

A Reddit post offers a rare glimpse into the approval process behind the open-source release of Qwen 3.7 models.