Mistral launches Voxtral TTS, completing its open-source speech pipeline
Mistral launched a text-to-speech model that clones a voice from 3 seconds of audio and runs on 3GB of RAM.
AI
The models, the launches, the funding rounds, and the quiet policy decisions that actually shape how artificial intelligence gets built and deployed.
-
-
Meta’s TRIBE v2 can predict how your brain responds to videos and sounds
Meta built a model that predicts how your brain responds to what you see and hear. They also own the world’s largest advertising platform.
-
Lovable adds AI penetration testing to its platform through Aikido partnership
Lovable just launched AI-powered penetration testing for $100. A traditional pentest costs up to $50,000. That gap deserves some scrutiny.
-
ComfyUI’s new Dynamic VRAM lets you run larger AI models with less RAM
RAM prices are up, AI models are getting larger, and your hardware is struggling to keep up. ComfyUI just shipped an update that helps with all three problems at once.
-
NVIDIA’s AutoGaze makes AI video understanding up to 19x faster by mimicking human attention
NVIDIA found a way to make AI video understanding up to 19x faster by teaching models to ignore the parts of a video that do not matter.
-
Firecrawl /interact endpoint allows you to chat with scraped web pages
Firecrawl launched a new endpoint that lets AI agents interact with a live browser session after scraping — logins, infinite scrolls, and dynamic content included.
-
Users are unhappy with Dreamina Seedance 2.0’s censorship and pricing
Seedance 2.0 launched to big hype, but the initial user reactions are far from positive.
-
Users report declining quality on Z.ai’s GLM-5 coding plans
GLM-5 arrived as one of the most compelling budget alternatives to frontier coding models. Users on Reddit are now reporting the quality is quietly getting worse.
-
xAI launches SuperGrok Lite at $10/month with basic image and video generation
xAI launched a $10/month SuperGrok Lite tier. The catch is that it puts capabilities that used to be free behind a paywall.
-
Google’s TurboQuant compresses AI memory by 6x without losing accuracy
Google just found a way to compress AI memory by 6x with zero quality loss — meaning longer conversations, faster responses, and cheaper inference on the same hardware.
