The models, labs, and research the AI world is actually talking about.
Developer posts video of his San Francisco run after announcing new role.
Former DeepMind researcher shares update on leaving Meta for new AI lab role.
The post comes from the account of X's owner and xAI founder.
Elon Musk states Grok operates independently on cloud infrastructure.
Perplexity AI enables tasks to shift from cloud to local models on Mac for private data.
Replies target an arXiv paper on sliding-window attention with sinks.

Grok is providing another reset on token usage limits for all its bot users.
OpenRouter says Mercury 2.5 Preview reaches 1,107 tokens per second exclusively on the platform with tunable reasoning and parallel tool calls.
Co-founders announce seed funding for an AI-powered physics research lab.
Investor shares details of second talk with Conviction founder on AI progress.
Team posts emphasize gains in accuracy, speed, and pricing for complex visual documents.
Wafer is an inference provider using Nvidia and non-Nvidia chips that declined buyout offers.
Tweet questions if people forgot rogue AIs can self-replicate without hacking AWS.
Former Tesla leader shares how he routes all Grok use through one controlling agent.
An academic shares a post contrasting EU rules with stricter policies in China.
Tenex co-founder Arman Hezarkhani posted his assessment of Notion on X.
Tweet details new level of game realism from generative models and real-time graphics.
The journalist shared a video of the company's unconventional launch parodying typical Silicon Valley announcements.
New mode shows fake code terminal while agents continue tasks in background.
Research engineer posts sarcastic X reply on feeding 50-item rubrics to small models.