True Positive Weekly #118
The most important artificial intelligence and machine learning news and articles
Hey, in this issue: the unreasonable effectiveness of an LLM agent loop with tool use; deeper insights into retrieval augmented generation; evaluating long-context question and answer systems; why video-language models can't see what humans can?; the bitter lesson is coming for tokenization; and more.
[Google] MedGemma: Google's most capable open models for health AI development
Time blindness: Why video-language models can't see what humans can?
[Model] Chatterbox: Leading open source voice cloning AI model
The unreasonable effectiveness of an LLM agent loop with tool use
[Tutorial] Evaluating long-context question and answer systems
[Google] Deeper insights into retrieval augmented generation: The role of sufficient context
[Tool] llm-d: A Kubernetes-native high-performance distributed LLM inference framework
Enjoy the newsletter? Please help us make it bigger and better by sharing it with colleagues and friends.