Machine Learning News: Chips, Robots, and Models - April 2024 Update

Explore the latest developments in machine learning, AI chips, robotics, and language models in this comprehensive video update. Dive into Meta's new training accelerator, Apple's data deal with Shutterstock, and OpenAI's internal challenges. Learn about advancements in language models like RecurrentGemma, CodeGemma, and Mistral AI. Discover breakthroughs in vision-language models, text-to-speech technology, and mobile UI understanding. Stay informed on government investments in AI and ethical considerations in AI development. Gain insights into cutting-edge research and practical applications shaping the future of artificial intelligence across various domains.

- Intro
- Our next-generation Meta Training and Inference Accelerator
- ALOHA Unleashed
- Apple Inks $50M Deal with Shutterstock for AI Training Data
- OpenAI Researchers, Including Ally of Sutskever, Fired for Alleged Leaking
- Adobe's Ethical Firefly AI was Trained on Midjourney Images
- Trudeau announces $2.4billion for AI-related investments
- RecurrentGemma: Moving Past Transformers for Efficient Open Language Models
- CodeGemma - an official Google release for code LLMs
- Mistral AI: Cheaper, Better, Faster, Stronger
- Vezora/Mistral-22B-v0.1
- WizardLM-2, next generation state-of-the-art-LLM
- Idefics2, the strongest Vision-Language-Model VLM below 10B!
- BlinkDL/rwkv-6-world
- Pile-T5: Trained T5 on the Pile
- Model Card for Zephyr 141B-A39B
- Parler TTS
- RHO-1: Not all tokens are what you need
- Ferret-UI: Grounded Mobile UI Understanding with Multimodal LLMs