MarkTechPost (2 hours ago)
Google Research released Agile Deliberation, a human-in-the-loop framework that helps users define subjective visual concepts through borderline images. In 18 live sessions, it lifted Gemini-based classifier F1 by 10.5 points over zero-shot.
Read the full story on MarkTechPost.
Experience cutting-edge VR and save up to 20% with coupons for the latest games, Meta Quest 3, Ray-Ban AI glasses, and more deals.
Delivering net zero also at risk, say MPs as they call for centralised policy and funding for mines and diversification The UK is “dangerously” dependent on China for critical raw materials needed for the military, renewable energy infrastructure and AI, a parliamentary investigation has found.
OpenAI on Wednesday detailed a pipeline that employed an upcoming cyber-focused artificial intelligence (AI) model to validate and patch 948 local privilege escalation vulnerabilities in Linux for $1.2 million in API costs.
OpenAI has rolled out ChatGPT Images 2.6 across consumer tiers, Codex, and its developer platform, introducing native animated GIF creation.
The company is said to be significantly dropping the amount it’s spending on AI for its engineering teams.
Meta VR Glasses are shaping up to be the mixed-reality wearables we wanted all along, and my test drive made me a believer.
webAI released TwIL-8B, an 8.8B reasoning model post-trained from IBM Granite 4.2 8B. It posts 98.2 on MATH-500 and 81.6 on LiveCodeBench, but trails similar-size models on knowledge benchmarks.
Ricursive Intelligence released Ripple-65B Preview, a 65B dense open-weight LLM under Apache 2.0. It scores 80.5 on Terminal-Bench 2.1, ahead of US open models up to 15x its size.
Tencent’s WeChat Vision team released WeVisOne-VL, an Apache 2.0 vision language model in 2B and 4B sizes. It handles detection, OCR, counting, segmentation and alpha matting in one model, with a 9.6M-parameter pixel decoder.
Anthropic released FlashProve, an Apache 2.0 zkSNARK prototype that proves a reply came from a committed model without revealing weights. It proves Llama 4 Maverick at 110 tokens/second on 1 H100, with 19 MB proofs.
69 active sources
24 Turkish · 45 English
All sources
