According to @grok on X, Grok now supports uploading complex PDFs and answering questions directly within the app and web, enabling retrieval augmented generation on long documents (source: Grok official post, Apr 1, 2026). As reported by Grok, users can query multi-section reports and technical papers, which suggests long-context parsing and semantic search to extract citations from large files. For businesses, this unlocks faster due diligence, policy compliance checks, and contract review by turning PDFs into interactive knowledge, according to Grok’s announcement. According to the same source, the feature is available in the Grok app and web, positioning Grok against ChatGPT’s Advanced Data Analysis and Claude’s attachments for enterprise workflows like RFP analysis and research synthesis.

According to AndrewYNg on X, DeepLearning.AI launched a short course titled "Agent Memory: Building Memory-Aware Agents," developed with Oracle and taught by Richmond Alake and Nacho Martínez, focused on persistent agent memory across sessions. As reported by DeepLearning.AI, the curriculum covers designing a Memory Manager for episodic, semantic, and procedural memory, implementing semantic tool retrieval to load only relevant tools at inference time without bloating context, and building write-back pipelines so agents autonomously update knowledge over time. According to the course page, the skills target production use cases like research agents that work over multiple days, enabling scalable retrieval, lower context costs, and improved task continuity for enterprise agents.

According to God of Prompt on X, new research comparing 9 memory systems across 1,540 questions finds retrieval methods, not write-time memory strategies, are the dominant driver of AI agent accuracy, with retrieval causing up to 20-point swings while write strategies yield only 3–8 points (as reported by the original X thread). According to the same source, raw conversation chunks with zero LLM preprocessing matched or outperformed fact extraction and summarization pipelines, indicating expensive preprocessing can discard useful context. The thread reports hybrid retrieval combining semantic search, keyword matching, and reranking cut failures roughly in half, and models used relevant context correctly 79% of the time, with retrieval quality correlating strongly with accuracy at r=0.98. For practitioners, this implies prioritizing hybrid retrieval, careful chunking, and reranking over token-heavy write-time compression to boost agent reliability and reduce costs (according to God of Prompt on X).

According to @AINewsOfficial_, semantic search technology now enables users to seamlessly browse and discover ERC-8004 AI agents across multiple blockchain networks using intent-driven queries. This innovation, showcased on 8004agents.ai, significantly improves accessibility to economic AI agents, streamlining user experience and lowering barriers for businesses to deploy and manage intelligent agents within decentralized Web3 ecosystems. The development opens new market opportunities for AI agent marketplaces, cross-chain interoperability, and intelligent automation solutions in blockchain-based economies (source: @AINewsOfficial_).