LangChainAI Introduces Multi-Agent Enterprise Research System
Today's curated insights on AI product management from X/Twitter across 60+ expert sources and YouTube channels.
LangChainAI Introduces Multi-Agent Enterprise Research System
From X
AI Product Launches & Updates
Enterprise Deep Research: LangChainAI @LangChainAI introduced a multi-agent system leveraging LangGraph for enterprise research automation, featuring real-time streaming and human-guided steering. Explore EDR on GitHub.
Chatsky Dialog Framework: LangChainAI @LangChainAI released Chatsky, a pure Python dialog framework with a dialog graph system integrated with LangGraph. Explore the framework.
Custom LLM Integration: LangChainAI @LangChainAI announced a production-ready solution to integrate private LLM APIs into LangChain and LangGraph apps, including authentication, logging, and state management. See implementation.
AI Tools & Applications
Perplexity Finance Sidebar: Aravind Srinivas @AravSrinivas added Perplexity Finance to the sidebar for easier daily access. View update.
ZenCity Orchestration Layer: Teresa Torres @ttorres showcased ZenCity’s AI-driven layer that reads public agendas and auto-generates council meeting prep packets tailored by department. Watch demo.
Product Management Insights & Strategies
Shipping as a Skill: Guillermo Rauch @rauchg emphasized that shipping includes design, QA, marketing, iteration, and that AI will push PMs to excel beyond coding. Read more.
Product Sense Defined: Aakash Gupta @aakashg0 highlighted that product sense is “the ability to find the right solution for the user and business despite limited and ambiguous information,” citing Sid Arora. Learn more.
Platform PM Toolkit: George @nurijanian outlined an 8-point toolkit for platform PMs, covering systems thinking, API design, DevOps fundamentals, data architecture, change management, and cross-team collaboration. See details.
AI Industry Developments & News
LLM Debugging Potential: Andrej Karpathy @karpathy shared a deep dive into a PyTorch MPS backend bug and speculated when an LLM might fully automate such debugging. Read thread.
Gemini App vs AI Studio: Logan Kilpatrick @OfficialLoganK explained the mental model distinguishing Gemini as a personal assistant and AI Studio as a builder platform, and hinted at integrating Gemini Apps Canvas. Full context.
From YouTube
AI Agents, Clearly Explained in 40 Minutes | Wade Foster (Zapier)
Peter Yang • October 26, 2025
Wade Foster and Peter Yang define the "AI automation spectrum"—from deterministic workflows to chat agents—demonstrate building a Zapier email-categorization agent using natural-language prompts, and share practical guidance on orchestrating specialized agents and driving AI adoption across teams.
Key Takeaways:
- Today's most reliable AI solutions sit in the middle of the spectrum—AI workflows and "agentic workflows"—offering balance between determinism, cost, and the reasoning power of LLMs.
- Zapier’s email-categorization agent uses natural-language instructions and integrations with Gmail, HubSpot, and archiving tools to classify over 100 emails daily into action-required, EA-handled, customer insight, and informational categories, reducing Wade Foster's inbox to under 10 priority messages.
- Effective company-wide AI uptake comes from hands-on hackathons, cross-department demos for knowledge sharing, and incremental trust-building by starting with non-destructive tasks (like labeling or drafting) before granting agents full automation rights.
How Block is becoming the most AI-native enterprise in the world | Dhanji R. Prasanna
Lennys Podcast • October 26, 2025
Dhanji R. Prasanna, CTO of Block, explains how he convinced leadership to adopt an “AI manifesto,” restructured Block into a functional org, and built the open-source AI agent Goose to drive productivity gains across engineering and non-technical teams.
Key Takeaways:
- AI-forward engineering teams at Block report saving 8–10 hours per week with tools like Goose, and company-wide metrics show a 20–25% reduction in manual work hours.
- Goose, built on the Model Context Protocol and open-sourced by Block, lets employees automate tasks across systems—from writing SQL in Snowflake to generating and emailing PDF reports—by composing AI “arms and legs.”
- Prasanna replaced Block’s GM-style siloed org with a unified functional structure—having all engineers and designers report to single tech leaders—to align strategy, share platforms, and accelerate AI-native development across Square, Cash App, Afterpay, and Title.