Intellectual hub of the topic
The Korean company Upstage has released the Solar Pro 3 language model, which handles multi-step agentic tasks twice as effectively as its predecessor.
PyTorch 2.11 has been released. This update to the popular neural network training framework brings notable improvements for distributed systems and Apple Silicon.
Fireworks AI has demonstrated that training AI models with reinforcement learning (RL) is significantly more affordable than the industry generally believes.
AI: Events
Technical context • Infrastructure
AMD has adapted the Miles framework for large-scale reinforcement learning on Instinct GPUs – now it works without NVIDIA hardware.
OpenAI has explained how the safety system in the Sora video generator and its associated application works – from moderation to digital credentials on content.
Researchers have introduced EvoClaw, an AI agent testing system that assesses the agents' ability to work with constantly evolving projects.
Nvidia has introduced OpenShell, a tool for safely running autonomous AI agents in a corporate environment. It isolates their actions and controls their access.
Fireworks AI explains why the race for megaclusters isn't the only path to powerful AI models and how reinforcement learning (RL) is changing the equation.
OpenAI has shared how it monitors deviations in the behavior of its internal code-writing AI agents and explained why this is crucial for safety.
Don’t miss a single experiment!
Subscribe to our Telegram channel —
we regularly post announcements of new books, articles, and interviews.