Tootfinder

DeepSeek-R1 erzeugt unsicheren Code bei politisch sensiblen Begriffen
Die chinesische KI DeepSeek-R1 erzeugt schlechteren Code, wenn Begriffe wie Falun Gong oder Taiwan im Prompt stehen. Das fanden Sicherheitsforscher heraus.

Nvidia says its GB200 Blackwell AI servers boost performance 10x compared to H200 servers for MoE models like Moonshot's Kimi K2 Thinking and DeepSeek's R1 (Stephen Nellis/Reuters)
https://www.reuters.com/world/china/nvidia

What a great read and overview, recommended !
"The State Of LLMs 2025: Progress, Problems, and Predictions"
#AI

The State Of LLMs 2025: Progress, Progress, and Predictions
A 2025 review of large language models, from DeepSeek R1 and RLVR to inference-time scaling, benchmarks, architectures, and predictions for 2026.

A year after the DeepSeek shock following R1's release, US tech companies still lead in AI chips, model quality, and sales, with no change in investment trends (Chris Miller/Barron's Online)
https://www.barrons.com/articles/deepseek-…

Demis Hassabis says Chinese AI companies are about six months behind leading western labs and the response to DeepSeek's R1 was a "massive overreaction" (Bloomberg)
https://www.bloomberg.com/news/articles/2026-0…