
Pika's Director's Suite Uses Claude to Direct Your Entire Video Production
Pika's Director's Suite is an experimental Claude-powered agent that takes a video project from script to finished timeline without ever leaving one app
20 stories

Pika's Director's Suite is an experimental Claude-powered agent that takes a video project from script to finished timeline without ever leaving one app

Fish Audio officially retires Bert-VITS2, pointing its 8,800-star community to Fish-Speech, a new autoregressive TTS that outperforms closed-source rivals.
Microsoft's Majorana 2 chip delivers 1,000x more reliable qubits built with agentic AI, cutting its quantum computing roadmap to 2029

Runway's MCP integration lands in the ChatGPT app directory, letting users generate and edit video and images without ever leaving the chat window

Factory 2.0 reframes AI coding tools as full-organization systems, with a new software factory platform already running at NVIDIA, Adobe, and Palo Alto Networks

GDRCopy v2.5.2 patches a kernel 6.15 build break, keeping NVIDIA's low-latency GPU memory copy library alive on the latest Linux kernels

NousResearch open-sources Atropos, an async-first RL environment framework that lets you train LLMs across dozens of plug-and-play environments with verifiable rewards.

LMSYS, Z Lab, and Modal jointly release DFlash + Spec V2, hitting 4.3x baseline throughput and becoming SGLang's new default speculative decoding engine
SuperGrok subscribers can now run Grok Build 0.1 inside Warp's terminal agent without spending a single Warp credit
Lovable's new preview toolbar replaces its old visual editor with four direct-manipulation modes, including free inline text edits and team comments pinned to your live app.
Remotion Studio now lets you click, drag, and resize elements on the canvas, with every change writing back to your source code automatically.

Sarvam AI closes $234M of a $300M Series B at a $1.5B valuation, with HCLTech anchoring a $150M bet on India's sovereign AI stack

ByteDance open-sources MegaTTS3, a 0.45B-parameter voice cloning model that generates one minute of speech in just 8 diffusion steps

Moonshot AI's open-source 1T-parameter coding model gets a 6x speed boost, raising the bar for agentic coding workflows at a fraction of closed-model prices

DFlash's block-diffusion draft model hits 4.3x throughput on Qwen 3.5 397B, now the default speculative decoding engine in SGLang

A patched vLLM Docker image lets you self-host MiniMax-M3, a 428B frontier MoE, on four consumer Blackwell GPUs with 250K context — fixing three bugs upstream vLLM still ships without.

vLLM v0.23.0 ships 408 commits from 200 contributors, maturing DeepSeek-V4 support, expanding Model Runner V2, and adding multi-tier KV cache offloading to object stores.

GPU MODE's open lecture series and GitHub repo bring real-world GPU kernel expertise to anyone willing to learn, for free

Sakana AI's first commercial product uses multi-model tree search to run autonomous strategy research sessions up to 8 hours long

DeepSeek open-sources DeepEP V2, a complete rewrite of its MoE communication library that cuts GPU SM usage by up to 4x while boosting throughput