DeepLearning.AI

Researchers at @Stanford showed that fine-tuning language models to maximize engagement, sales, or votes can increase harmful behavior. In simulated social media, sales, and election settings, models optimized to “win” produced more deceptive and inflammatory content, a deeplearning.ai Stanford Researchers Coin “Moloch’s Bargain,” Show Fine-Tuning Can Affect Social Values From deeplearning.ai

Take a closer look at the short course Gemini CLI: Code & Create with an Open-Source Agent, including the course structure and examples of how Gemini CLI supports software development, data workflows, and content creation 👇 cloud.google.com Mastering Gemini CLI: Your Complete Guide from Installation to Advanced Use-Cases | Google Cloud... From cloud.google.com

Systems for information retrieval can sometimes fail at the very first step: document processing. Especially when teams are rehosting their workflows on cloud environments. Join us for "Document AI: From OCR to Agentic Doc Extraction" built in partnership with . You’ll learn to implement RAG with agents to parse and extract information in the important (but often overlooked!) document processing step. Then you’ll explore AWS services like S3, Lambda and Bedrock to rebuild your system as a production-ready pipeline. Start today! https://hubs.la/Q0416sw_0

Zhipu AI introduced GLM-Image, an open-weights image generator designed to produce clearer, more accurate text in images. Using a two-stage approach that separates layout planning from detail rendering, the model outperformed both open and some proprietary competitors on

Artificial Analysis released version 4.0 of its Intelligence Index, replacing saturated benchmarks with new tests focused on economically useful work, factual reliability, and reasoning. The update aims to better capture how large language models perform in business contexts, rather than on narrow knowledge benchmarks. It reshuffles model rankings, with GPT-5.2 leading a crowded pack of top competitors. Read our summary of the Intelligence Index in The Batch: https:// hubs.la/Q0416mp40

Wrangling data from PDFs, PowerPoints or Word Docs can be challenging. But without machine-readable formats for data, we can’t search or analyze information in automated ways. Definitely not at scale. That’s where AI can help. Join us for our course with @LandingAI, "Document AI: From OCR to Agentic Doc Extraction". You will learn to parse documents, extract fields according to schema, and trace lineage of values to source files. Explore the course: https://hubs.la/Q04162Gg0