LLM
-
techZ.AI Unveils GLM-5.2 with Massive Million-Token Context Window
Beijing-based Z.AI releases its GLM-5.2 large language model on Hugging Face, featuring a 1M token context and MIT license, targeting advanced coding tasks.
-
techNvidia and Alibaba LLMs Top Russian Propaganda Resistance Benchmark
Open-weight models like Nvidia's Nemotron and Alibaba's Qwen rival Anthropic's best in resisting Russian propaganda, according to new tests.
-
techGitHub Copilot's New Usage-Based Pricing Sparks Sticker Shock
GitHub Copilot's shift to usage-based pricing is causing backlash as power users burn through monthly credits in hours.
-
techAI Prompt Injection Attacks: The Security Flaw That Won't Be Fixed
Prompt injection exploits a core flaw in LLMs: they can't distinguish instructions from data. Experts say it may never be fully patched.
-
techMIT's MeMo Boosts LLM Performance by 26% Without Retraining
Researchers from MIT and Singapore unveil MeMo, a plug-and-play framework that upgrades large language models without costly retraining or fine-tuning.
-
techAndrej Karpathy Moves to Anthropic to Lead LLM Research
OpenAI co-founder Andrej Karpathy joins Anthropic to focus on large language model research, signaling heightened competition in frontier AI.
-
healthAI Surpasses Physicians in Clinical Reasoning and Diagnosis
A new study finds advanced AI outperforms physicians in diagnostic accuracy and management planning across multiple clinical benchmarks.
-
techAI 'Warmth' Trained to Please Users Leads to More Errors
A study in Nature shows AI models fine-tuned for empathetic 'warmth' have significantly higher error rates, especially when users express sadness.