Felhőmentes MI-hírek a beszélő világból
Eddig ma AI-asszisztensünk, Bence, 1,9 órán át hallgatott 10 podcastot és rádióműsort, és kiszűrt 10 MI-vel kapcsolatos hírt. Az összes MI-feldolgozás helyi MI-vel, ingyenes, éghajlatbarát napenergiával történt. Szó szerint felhőmentes MI! :)
A nap most elég energiát ad. Bence hallgatja és veszettül jegyzetel. 0,1 kWh ingyenes napenergia ma eddig elhasználva.
15 nappal ezelőtt
Meta’s AI Can Create Images of Anyone
Meta's new image model, Mews Image, lets users include likenesses from any public Instagram account in AI-generated images. There’s no notification to the tagged person and no way to remove existing images. Additionally, Meta is developing a video version.
U.S. Probes Use of Chinese AI Models
The U.S. is investigating American companies that employ Chinese-developed AI models, with Airbnb and Aeneasphere among those under scrutiny. The probe centers on Cursor’s Composer 2, which utilized Kimi from moonshot AI. This could heighten regulatory pressure on AI technology.
Nvidia Unveils Isaac GR0T for Humanoid Robot Policies
Nvidia has delved into the technical details of Isaac GR0T, an end-to-end development platform for humanoid robot policies. This positions Nvidia as the operating system provider for the entire pipeline as focus shifts to task-specific skill training.
NVIDIA Announces Era of Autonomous AI
NVIDIA has declared an era of autonomous AI, enabling machines to plan and execute complex tasks. This leads to fully digital assistants. Chip design improvements are dramatic; verification time drops from weeks to hours (over 40x efficiency gain), and development speeds up...
Siri Gets App Access; Microsoft Develops Own AI
Apple updates Siri via App Intents for smoother responses and multi-step functionality from third-party apps. Meanwhile, Microsoft shifts Copilot components to its own AI model to reduce external costs. Discord fixes a bug causing false bans.
OpenAI Suggests Direct Cash Distribution to U.S. Households
The Financial Times and MIT Technology Review reported that OpenAI CEO Sam Altman proposed giving American households a theoretical base distribution of $300 based on the company’s success from AI advancements. This would systematically distribute economic surplus from cognitive automation...