the-decoder2d
Researchers create massive multimodal dataset to train a tiny AI model that beats the big ones
A team of researchers from various Chinese institutions has created Infinity-MM, one of the largest publicly available datasets for multimodal AI models, and trained a new top-performing model on it.
the-decoder2d
OpenAI hires Meta's AR glasses chief to lead new robotics and hardware push
Caitlin Kalinowski, the former head of Meta's AR glasses development, is joining OpenAI. The AI specialist is apparently planning a major push into robotics and consumer hardware with this move. As ...
the-decoder3d
Anthropic releases updated, smarter Claude Haiku 3.5 and Sonnet 3.5 model
Anthropic has released its new Haiku 3.5 model, now available through the Anthropic API, Amazon Bedrock, and Google Cloud's Vertex AI. According to Anthropic, the model shows improved abilities in ...
the-decoder3d
AI-generated game Oasis now turns images into playable 3D worlds
Decart AI has released an update to its Oasis AI game generator that allows users to upload their own source images for the AI to turn into playable game worlds. The new feature, called "Custom Worlds ...
the-decoder3d
Netflix creates new leadership role focused on AI game development
Mike Verdu, who has led Netflix's games division for the past three years, will become vice president of GenAI for games. Verdu announced on LinkedIn that he plans to adopt a "creator-first vision […] ...
the-decoder3d
X's Grok chatbot adds vision capabilities
Musk called this an "early version" of the feature and said it would "rapidly improve." Other chatbots and their underlying models, such as ChatGPT, Google Gemini, and Meta AI, have had this visual ...
the-decoder3d
Polish radio station ends AI presenter test after public backlash
A Polish radio station stopped using AI-generated radio hosts after just one week following strong public opposition to the three-month project. OFF Radio Krakow, a public radio station, halted its AI ...
the-decoder4d
How to turn off ChatGPT's default web search
While you can force a web search using a small web button in the chat window, turning it off takes a few extra steps. If you want to stop seeing internet sources with your responses, here's how to ...
the-decoder4d
Google adds search to Gemini API with premium pricing
Google announced that developers can now integrate Google Search into the Gemini API, a feature called "Grounding with Google Search." The integration aims to provide more accurate and up-to-date AI ...
the-decoder4d
Open-source Moonshine speech recognition model is up to five times faster than OpenAI's Whisper
U.S. startup Useful Sensors has developed Moonshine, an open-source speech recognition model that processes audio more efficiently than OpenAI's Whisper while using fewer computing resources. The ...
the-decoder4d
New AI web navigation system uses world models to predict outcomes and boost success rates
Researchers at Korea's Yonsei University have created a new AI system that tests actions before executing them on websites. The approach shows better results than previous methods at helping AI ...
the-decoder4d
Meta's new robotic fingertip mimics "human-level multimodal sensing capabilities"
Meta FAIR has developed three new technologies to give robots better touch capabilities. The company plans to make these technologies available to researchers in 2024. The key innovation is Meta Digit ...