Browse every post or search by title and content.
HuggingFace and ServiceNow release EVA-Bench Data 2.0, a massive enterprise AI agent benchmark with 121 tools and 213 scenarios across IT, H...
GitHub Universe 2026 returns to San Francisco with a focus on the agentic era, where AI assistants evolve from autocomplete to autonomous de...
Vercel updates its Terms of Service to clarify shared liability when AI agents take autonomous actions on user accounts, setting a precedent...
NVIDIA and Hugging Face released Nemotron 3.5, an open-source multimodal safety suite for content moderation. It supports fine-tuning for cu...
If you searched "anthropic api 529 error," there's a good chance you're staring at something like this in your terminal right now:⎿ API Erro...
New ontology-grounded framework for pre-deployment AI agent verification aims to close the gap between LLM benchmarking and production safet...
Nvidia's Task-Seeded Synthetic Data Generation creates high-quality task-specific training examples, reducing dependency on human annotation...
A new arXiv study finds encoder-only Transformers beat LSTMs for streamflow prediction in ungauged basins, achieving 18% lower error and mai...
New research shows DPO works for code generation, search ranking, and robotics—not just chatbots. Reduce RLHF costs by 40% with a single uni...
xAI's Grok Imagine Video 1.5, now live on Vercel AI Gateway, generates video and synchronized audio from a single image. The model improves ...
New research shows LLMs can use internal graph scaffolds to organize reasoning, reducing reliance on external knowledge bases. Early tests s...
Vercel Blob now supports signed URLs with per-operation scoping and 7-day expiry, enabling secure direct uploads and delegated access withou...