Cursor Publishes Post-Mortem on Recent Security Incident
Cursor published a detailed overview of a recent security incident along with the remediation steps it has taken.
Global AI Digest
-
-
LiteLLM Launches Customer Portal for Usage-Based AI Budgets
LiteLLM released a new Customer Portal that helps teams set budgets for customers, cap spending accurately, and load balance between them for usage-based AI pro…
-
Jev Introduces "Speculative Questions" Batching Pattern
TypeSafe AI shares a Jev best practice: batching questions into a single call even when you don't yet know if you'll need the answers, because it is often faste…
-
Gemini 4 Hydrogen Turns a Simple Brief Into a Finished Video
Gemini 4 Hydrogen can take a short brief and a custom skill to produce a finished video, sourcing public-domain archival footage and assembling it into a single…
-
Tencent's Hy4 Tops CWE-Bench-AA at 76%
Tencent's Hy4 ranks first on CWE-Bench-AA with 76%, six points ahead of DeepSeek V4.1 Flash 1017 (max).
-
Hugging Face Highlights End-to-End Transformers Export, Quantization, and Cross-Backend Execution
A shared thread describes exporting, quantizing, and running every transformers model end-to-end on any backend with identical inputs and outputs as a drop-in r…
-
Jev Achieves 94ms per Photo, ~30x Faster Than Sol at ~50x Lower Cost
TypeSafe AI's Jev processes a photo in 94ms, roughly 30x faster than Sol, and delivers about 50x as many judgments for the same cost.
-
Anthropic Publishes 10 Recommendations for Designing and Reporting AI Safety Evaluations
Anthropic's Engineering Blog shares 10 recommendations for how to design, run, and report safety evaluations that determine whether a model is ready to deploy.
-
TRL v1.18.0 Stabilizes AsyncGRPOTrainer for Parallel Generation and Training
TRL v1.18.0 makes AsyncGRPOTrainer stable, running generation and training simultaneously on separate GPUs instead of alternating, with documentation on usage a…
-
AWS Publishes Guide to Building a Self-Service LLM Gateway with Bedrock and LiteLLM
An AWS team published a walkthrough on building a self-service LLM gateway using Amazon Bedrock and LiteLLM.
