TLDR AI 2024-09-18

Don't wait for the AI future—come build it.

This September, join the world's largest gathering of open source AI infrastructure leaders and an all-star lineup of speakers you won't see anywhere else.

Attend Ray Summit 2024 in person to:

→ Go behind the scenes of some of the most ambitious and large-scale AI powered consumer applications.

→ Mingle with the world's top minds in distributed computing, machine learning, and AI.

→ Explore the Ray and Anyscale roadmaps and meet the teams that lead them.

🌉 Register to attend @ The San Francisco Marriott Marquis

Mistral has released a free API tier, dramatically reduced its costs, improved the performance of its small model, and put its vision model in Le Chat.

Nvidia's dominance in AI chips has propelled it to immense market value, largely thanks to its GPU capabilities and CUDA software ecosystem. However, competitors like AMD, Intel, Cerebras, and SambaNova are developing innovative solutions to challenge Nvidia's supremacy in AI hardware. While Nvidia's lead remains secure for now, the landscape is dynamic, with multiple players striving to carve out their own niches in the AI market.

The Jina series of embeddings are a high quality and powerful suite of models that can be used for embedding and retrieval. Its development team has released the next version of their model with improved performance and training.

This work introduces a framework to evaluate the trustworthiness of Retrieval-Augmented Generation (RAG) systems across six key areas: factuality, robustness, fairness, transparency, accountability, and privacy.

The new beeFormer framework improves sentence Transformers by incorporating interaction data, making them more effective for recommender systems.

The last frontier of Visual Language Models is the ability to understand and reason about comics. This project is a survey and a call for research.

Join Bloomreach, Google, and NVIDIA for this exclusive one-day event! Hear from leaders across industries about how they're bringing AI to life, where they see opportunity, and their predictions for what's next. Save your virtual front row seat to tap into the future of AI in ecommerce - told by the people building it day by day.

A lightweight toolkit for fuzzy deduplication, reranking, and other NLP-based tasks. Optimized to run on the CPU.

This project enhances speech representation learning by separating syllabic structures from speaker information in self-supervised models. By fine-tuning the HuBERT model with speaker perturbation techniques, researchers improved syllable segmentation, leading to better syllabic unit organization.

AI tools like GitHub Copilot enhance programming productivity but risk eroding essential coding skills. Over-reliance on AI-generated code can lead to quality, security, and maintainability issues and reduce learning opportunities. These tools may also limit creative problem-solving and foster a false sense of expertise among developers.

With models being somewhat commoditized, much of the advantage in AI comes from the data. It also, by extension, comes from the pipeline that ingests and creates the data. This post discusses the challenges and opportunities associated with data pipelines in the modern age.

ByteDance is developing its own AI chips with TSMC to reduce reliance on Nvidia GPUs amid U.S. export controls.

A custom-trained Florence 2-based model and system that can summarize CCTV and Surveillance videos and give accurate updates about what is occurring at any time.

Salesforce has debuted Agentforce, its effort to create generative AI bots capable of taking action on their own - within established limits.