A deep dive into how the 'Gumbel watermarking' method embeds an invisible trace into AI text – and why it's so tricky to find.
AI: Events
Illustrious XL 3.5: When an Image Generator Starts Understanding Language Like a Language Model
Products
Illustrious XL has been updated to versions 3.0–3.5: the new model supports resolutions up to 2048 pixels and understands complex text prompts on par with small language models (LLMs).
AI: Events
Why Voice Recognition Fails in Hospitals and How AI is Learning to Speak Medicine
Medicine
Standard speech recognition systems struggle in medical settings. We explore why this happens and what specialized AI has to offer.
The Sarvam AI team conducted a large-scale study on the quality of speech recognition systems for Indian languages, highlighting the challenges they uncovered.
LG Research has introduced EXAONE 4.5, an open multimodal language model capable of simultaneously analyzing text and images.
Alibaba introduces Qwen3.6-Plus, an enterprise AI model capable of independently developing code and analyzing visual content in real-world scenarios.
AI: Events
Alibaba Unveils Wan2.7-Image: Precise Color, Lifelike Characters, and Error-Free Text
Products
Alibaba has unveiled Wan2.7-Image, a unified image generation model featuring precise color control, personalized characters, and support for 12 languages.
AI: Events
Falcon Perception: A Single Transformer for Vision, Text, and Document Understanding
Technical context • Research
The Falcon Vision team from the Technology Innovation Institute has introduced Falcon Perception, a compact model that can identify and highlight objects in images based on text descriptions.
Google has released Gemini 1.5 Flash Live – an updated model for voice interaction with AI that has become more natural and reliable in real-world scenarios.