How AI truly advanced in 2025: Andrej Karpathy highlights 3 key points

When a blog post by Andrej Karpathy lands in your feed, you pay close attention, simply because few voices in the field of ...
InfoQ

Meta Details GEM Ads Model Using LLM-Scale Training, Hybrid Parallelism, and Knowledge Transfer

Meta released details about its Generative Ads Model (GEM), a foundation model designed to improve ads recommendation across ...

Anthropic reveals that as few as '250 malicious documents' are all it takes to poison an LLM's training data, regardless of model size

Claude-creator Anthropic has found that it's actually easier to 'poison' Large Language Models than previously thought. In a ...
MIT Technology Review

OpenAI has trained its LLM to confess to bad behavior

Large language models often lie and cheat. We can’t stop that—but we can make them own up. OpenAI is testing another new way to expose the complicated processes at work inside large language models.
Yahoo Finance

Cerebras Delivers End-to-End Training and Inference for Jais 2, the World’s Leading Open Arabic LLM

Trained on the industry’s largest, highest-quality Arabic-first dataset, Jais 2 sets new standards for accuracy, fluency, and cultural intelligence SUNNYVALE, Calif., December 09, 2025--(BUSINESS WIRE ...

Korean AI startup Motif reveals 4 big lessons for training enterprise LLMs

Motif-2-12.7B-Reasoning is positioned as competitive with much larger models, but its real value lies in the transparency of ...
Hosted on MSN

Kyivstar selects Gemma as base model for training national LLM

Kyivstar (KYIV) and the Winwin AI Center of Excellence under the Ministry of Digital Transformation of Ukraine have selected Google’s (GOOG) Gemma as the base model for training the large language ...
Hackaday

It Only Takes A Handful Of Samples To Poison Any Size LLM, Anthropic Finds

It stands to reason that if you have access to an LLM’s training data, you can influence what’s coming out the other end of ...