Multimodal AI
Multimodal AI breakthroughs and products
Articles
The Download: How People Really Use AI, and Flock’s Design Choices⭐8
New research from the AI Observatory reveals how people actually use AI, uncovering sensitive behaviors hidden in major companies' reports.
AI's Recursive Self-Improvement May Not Arrive as Fast as Predicted⭐8
AI's recursive self-improvement may arrive slower than expected due to creativity limits and research gaps, despite industry optimism.
We Still Don’t Know How People Are Really Using AI⭐10
New study reveals AI use is far more personal than work-focused, challenging tech giants' productivity claims with surprising insights.
Xemo-Talker: Unlock Emotions Explicitly for Audio-Driven Talking⭐9
Xemo-Talker enables precise emotion control in audio-driven talking portraits via less-principal subspace supervision, balancing lip sync and expression quality...
What Flock's Defenders Overlook in the Surveillance Debate⭐9
Flock Safety defenders overlook key design choices in surveillance debates, from data retention to cross-jurisdiction sharing and privacy risks.
The Download: When Robot Friends Die and the Rise of the⭐7
This is today's edition of The Download , our weekday newsletter that delivers a daily dose of what's happening in the world of technology.
What Happens When a Kid's Robot Best Friend Dies?⭐9
The Fragile Bond of AI Toy Companions: What happens when a child's robot best friend breaks or the company dies? Exploring the emotional fallout.
How Much Hydrogen Awaits Us Underground?⭐8
Global race for underground hydrogen heats up, but estimates vary. Explore how much natural hydrogen may await and the challenges of tapping this clean fuel sou...
Can Vision-Language Models Assess Proxemic Risk from Egocentric⭐10
Current VLMs show limits in egocentric proxemic risk assessment, though targeted prompts and fine-tuning boost high-danger recall in models like Qwen-VL.
HIMEC: Directional Change Representation and Fixed-Interface⭐9
HIMEC introduces directional change representation and fixed-interface decoding for remote sensing image change captioning, achieving superior CIDEr