Prime Intellect

नवीनतम AI समाचार, मॉडल और रिलीज़ Prime Intellect.

शोध 🇺🇸

Import AI 457: Stuxnet-like AI, cursed Muon optimizer, and positive alignment

SentinelOne researchers discovered a 20-year-old computer virus called fast16 that selectively distorts results of high-precision calculations, similar to futuristic weapons against scientific progress. Tilde Research identified a serious flaw in the Muon optimizer—it kills neurons in MLP layers—and proposed an alternative, Aurora. A group of scientists introduced the concept of 'positive alignment' for AI, aiming not only for safety but also for actively promoting human flourishing.

SentinelOne

Prime Intellect

OpenAI

Anthropic

Google DeepMind

Import AI27.07 · 12:06

मॉडल 🇺🇸

Nemotron Labs: How Open Models Give Enterprises and Countries AI They Can Trust, Govern, and Customize

Open models such as NVIDIA Nemotron enable enterprises and governments to build specialized AI with full control and customization. Examples from Abridge, Glean, Harvey, and others show how customization improves accuracy and reduces costs compared to closed models.