Data Points

OpenAI introduces new canvas UI for editing with ChatGPT. Google’s chip design model gets a new (but familiar) name. Microsoft launches some of the features it promised with Recall. Aider wants to give the right models the right jobs.

Hacking ChatGPT’s long-term memory function. U.S. trade commission targets companies who lie about AI. A new OpenAI model for screening text and images. Apple, Meta, and others hold off on AI Pact

Google’s top Gemini models cut prices, boost performance. Microsoft’s new approach to correct hallucinations. OpenAI releases a multilingual training dataset. Chain-of-thought reasoning has limits.

Kling 1.5 text-to-video model adds 1080p, editing tools. Codeforces restricts AI use in competition. Identifying whalesong using bioacoustic markers. A prize challenge for improving robotics world models.

Runway adds video-to-video and API. Moshi, a new open speech model. LlamaCoder’s open webapp builder alternative. California restricts synthetic actors and election deepfakes.

Hugging Face open-sources an LLM evaluation suite, Adobe announces its Firefly Video model, Meta researchers blend image diffusion with text transformers, NotebookLM can now generate synthetic podcasts.

Copilot adds fine-tuning for faster code completion, DataGemma uses RAG and RIG for fact-retrieval, Mistral introduces its open multimodal model, results of the latest summit on military AI.

NVIDIA’s Blackwell chips impress on hardware tests, fine-tuned versions of Llama 3.1 add reflection, most AI jailbreaks may not amount to much, new architecture extends context windows to 100M tokens.

AlphaProteo, a DeepMind system that designs novel proteins, updates and price drops for Command-R and Command-R+, Anthropic shows off easy software projects in Claude, YouTube builds system to detect synthetic music and faces.

The Open Source Initiative updated its Open Source AI Definition draft, clarifying that both AI models and weights must meet open source standards.

Qwen2-VL is available as 2 billion and 7 billion parameter versions under an Apache 2.0 license, as well as a 72 billion parameter API version.

NVIDIA and Mistral AI introduced Mistral-NeMo-Minitron 8B, a new language model that outperforms similarly sized models on multiple benchmarks.

The two new Jamba models use a hybrid solid state model-transformer architecture, feature an effective 256,000-token context window, and outperform competitors in their size

ByteDance’s China-only video app, U.K. watchdog probes Anthropic’s deals, New Writer models for medicine and finance, SambaNova’s speedy inference platform

A new automated paper-writing system from Sakana AI, GitHub’s Autofix uses GPT-4 to find and plug security holes, Google’s new AI voice assistant...