Author(s): Towards AI Editorial Team

Originally published on Towards AI.

This week has been busy with major AI developments, as is becoming the norm, dominated by new model releases and lawsuits. It was a difficult week for OpenAI in particular. Elon Musk sued OpenAI and CEO Sam Altman after a long-running disagreement about OpenAI’s departure from its original charity structure and open-source focus. Anthropic also released its new Claude 3 series of models, which now beats GPT-4 on most of its evaluations. The Stable Diffusion 3 research paper was also released, revealing more details about its architecture — a diffusion transformer similar to OpenAI’s Sora.

LLM benchmark tests get much criticism given how easy they can be to game (or accidentally include in the training data) and how they differ from real-world experience and use cases. Nevertheless, achieving leading scores is an impressive achievement. GPT-4 finally has competition not only from Gemini 1.0 Ultra and Gemini 1.5 Pro but also Claude 3 Opus. Particularly positive, however, is that Claude 3 has been made immediately available to test, and real-world feedback has been very positive, with many people surprised at its abilities, from science understanding to coding and literature.

While Claude 3 is not necessarily a step change in LLM capability, it seems… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI