Upgrade the Thousand Questions Big Model to a Real-Time Speech Recognition Big Model (Fun-ASR-Realtime)
Related
MIRA, an open-source 5B multiplayer world model, uses DINOv3 representation to mitigate long-term drift and can simulate 2v2 battles in Rocket League in real time.
According to Beating, AI research firm General Intuition, in collaboration with French AI lab Kyutai and Epic Games, has launched MIRA, a multiplayer interactive world model. As a generative game simulator supporting real-time multiplayer interaction, MIRA can simulate 2v2 battles in Rocket League in real time, based solely on historical footage and player button presses, without requiring a physics engine, rendering engine, or explicit 3D representation for inference. Unlike the "decoupling of logic computation and image rendering" approach adopted by companies like Odyssey, MIRA takes a generative simulation approach based on video latent space. MIRA boasts 5 billion parameters, and its core design builds the latent prediction space on a frozen general-purpose visual encoder, DINOv3-L. Leveraging pre-trained visual features, the generated latent states can more stably fall within the effective representation space, significantly mitigating image drift and divergence during long-term prediction. For multi-screen alignment, MIRA stitches the latent images from four player perspectives into a unified grid, enabling spatial attention mechanisms to operate naturally across viewpoints and improving the spatial consistency of vehicles, the soccer ball, and key events across multiple perspectives. The Action Dropout introduced during training also helps the system complete the game behavior of vehicles not controlled by commands when parts of the motion flow are missing. Currently, MIRA can run in real time at 20 frames per second on a single NVIDIA B200 graphics card. The team has open-sourced the training and inference code and released the Rocket Science dataset, which contains 1,000 hours of matches, approximately 4,000 hours of video, motion flow, and physics data from four perspectives; the complete training of the model used approximately 10,000 hours of clean match data.
AlphAi upgrades Polymarket's compatibility with prediction market experiences, adding AI analytics and real-time signals.
According to official Odaily, Web3 trading platform AlphAi has announced an upgrade to its Polymarket-compatible prediction market experience, further integrating prediction markets into the platform's trading ecosystem. This upgrade adds AI analytics, real-time market signals, structured navigation, and Crypto market categorization. Users can view AI reference analysis, social dynamics, Smart Money activities, and real-time event markers on supported prediction market pages to better understand the information flow, fund movements, and event progression behind changes in market probability. AlphAi stated that prediction markets are gradually evolving from simple order book access to a more intelligent, event-driven trading experience. The platform will continue to build a trading discovery and entry point for Web3 users, focusing on memes, prediction markets, crypto events, and other emerging trading scenarios.
Shengshu Technology Releases New Real-Time Interactive Model Vidu S1
Mars Finance reported on July 3rd that at the 2026 Global Digital Economy Conference's Artificial Intelligence Integration and Application Development Forum, Zhu Jun, founder of Shengshu Technology, officially released the Vidu S1 Real-Time Interactive Model, a new generation model designed for real-time interactive scenarios. The model supports real-time video calls and voice control of video feeds, allowing users to not only control the behavior of digital humans through voice but also achieve continuous interaction for an unlimited duration. Simultaneously, Vidu S1 supports 540P (960x540) high-definition resolution and a 25FPS frame rate (supporting up to 42FPS). It can quickly create exclusive interactive characters based on any initial image and personalized voice, including real people, animations, and cute pets, providing users with a more natural, smooth, and immersive real-time interactive experience.
StandX has upgraded its Maker Uptime Program to SIP-5A, offering daily DUSD earnings and platform token rewards.
According to Foresight News, decentralized perpetual contract exchange StandX tweeted that its original Maker Uptime Program has completed its June 2026 reward distribution and has been officially upgraded to SIP-5A: Community Maker Yield. The upgraded SIP-5A will provide market makers with daily DUSD (real stablecoin yield) and StandX platform token rewards. Users only need to provide liquidity on the StandX order book (placing both buy and sell orders) to participate in the program; no funds need to be locked. Existing Maker Hours and rewards will be automatically continued.
Vitalik specifically praised the Thousand Questions model for its outstanding anonymous identification capabilities.
ChainCatcher reports that Vitalik Buterin has announced a winner in his AI anonymization challenge. In the challenge, he wrote a revised version of EIP-7503 (Zero-Knowledge Wormhole) in Chinese in 2024, manually correcting it after using Qwen 2.5 for local translation, attempting to conceal his authorship. The article states that although the essay style was effectively obscured, the AI successfully identified the author by detecting unique thought processes in mathematical and algorithmic explanations (such as concrete numerical examples and logical chains), completely bypassing its obfuscation strategy, which was only applied to essays.
Claude Code's real-time collaboration feature is available to individual users, while Pro and Max subscriptions can generate real-time web dashboards.
According to Beating, Anthropic announced that it will roll out the Live Artifacts feature of its terminal agent programming tool, Claude Code, to Pro and Max individual subscription accounts. Previously, this ability to convert command-line interactions into web dashboards in real time was only available to team and enterprise users. Now, individual developers can directly issue commands in the terminal when writing code or analyzing data using Claude Code, publishing session-generated data, prototypes, or architecture diagrams as independent web pages. The generated web pages will refresh in real time as the terminal task is executed. By sharing a private link, team members or customers can view the latest progress at any time without developers needing to repeatedly send screenshots or share terminal screens. Currently, these web pages are statically hosted, unable to send network requests to external networks from within the page, nor can they save form inputs or perform multi-route redirects. Users can use this feature after logging in to the terminal.