Today’s Story

Meta’s Chameleon AI Bridges the Gap between Open Source and Commercial models.

Chameleon is a new family of AI models from Meta that can understand and generate both images and text, as well as process combinations of the two modalities:

-The model comes in 7 billion and 34 billion parameter versions, demonstrating strong performance across image captioning, text-only tasks, and non-trivial image generation.

Compared to Industry Leaders: While not yet at the level of more prominent AI models like Google’s Gemini Pro or OpenAI’s GPT-4V, Chameleon represents a significant step forward for open source multimodal AI.

The researchers behind Chameleon claim to have made substantial progress since the initial training of the models five months ago, hinting at even more advanced capabilities in the near future.

The race between open and closed source models has a myriad of implications. Open source models, fundamentally, are less regulated and controlled by their developers therefore coming with safety concerns, while at the same time carrying less platform risk. It remains to be seen what kinds of models will flourish in the coming months.

Headlines

• Daydream $50M Round to Build AI E-Commerce Search Engine (TC)- They will be focused on personalized shopping in an effort to optimize search once again for online retailers.

• The Future of Warfare Shaped by Chips (COAI) - “The computer itself was born out of the necessities of war, with early machines like Colossus being built to crack enemy codes during World War II”

• PayPal Hires Walmart Exec in AI Push (RT) - “One of Walmart's top tech executives Srini Venkatesan to head up the company's push into artificial intelligence as its new chief technology officer”

Research

Anthropic’s Claude 3.5 “Sonnet” Boosts AI Capabilities

• Anthropic releases Claude 3.5 “Sonnet,” enhancing the AI model’s capabilities in vision, understanding humor , and performance, while expanding its use cases through the new “Artifacts” feature.’

• Claude 3.5 Sonnet outperforms its predecessor, Claude 3 Opus, on various benchmarks, including graduate-level reasoning questions and undergraduate-level knowledge, while also being faster and more cost-effective.

• The model’s vision capabilities have also been greatly enhanced, enabling it to accurately transcribe text from imperfect images and interpret charts and graphs more effectively than before.

• Sonnet demonstrates a better grasp of humor and sarcasm, making interactions feel more natural and human-like.

Tool Spotlight

AI Creations

Each Day we Highlight Gen AI Art and Projects!

The COAI Podcast