Deep agents are a form of agents that plan, reason, and act over longer time horizons. We built a dedicated UI for viewing and interacting with these agents that show its plan, the status of the file system that it uses, and any sub aents it kicks off. My name is Nick.I'm an engineer at Langchain, and today you'll learn how to set up this UI. Now, as a quick refresher, we can think of deep agents as a variant of the generic React tool calling architecture. Under the hood, deep agents still follow the same i ...

RT Marius | Kamino (@y2kappa)Open source, formally verified yieldOpen source, formally verified borrowing ...

“The 'Open' in OpenAI is supposed to mean 'open source'. And it was created as a nonprofit open source. And now it is a closed-source for maximum profit, which I think is not good karma.”Elon Muskhttps://t.co/Dn9eo3HFPx ...

Core Functionality - LlamaIndex introduces FlowMaker, an experimental open-source visual agent builder enabling AI agent creation via drag-and-drop without coding [1] - FlowMaker automatically generates TypeScript code for visual flows [1] - The platform integrates with LlamaCloud indexes and tools [1] - It offers an interactive browser testing environment for real-time feedback [1] Key Features - FlowMaker features a visual drag-and-drop interface for no-code agent development [1] - It supports complex flow patterns with loops and conditional logic [1] Use Cases - FlowMaker facilitates basic agent creation by connecting user input nodes to language models [1] - It enables tool integration, demonstrated by a resume-searching agent using LlamaCloud indexes [1] - The platform allows implementing decision logic, conditional branching, and loop-back mechanisms for intelligent conversation routing [1] Feedback - LlamaIndex is actively seeking user feedback on FlowMaker [1]

Agent Architecture & Functionality - The Langchain deep research agent is highly configurable and open source, allowing for customization to specific use cases [1] - The agent operates in three main phases: scoping the problem, research, and report writing [3] - The research phase utilizes a supervisor to delegate tasks to sub-agents for in-depth research on specific subtopics [4] - Sub-agents use a tool calling loop, which can be configured with default or custom tools (like MCP servers) for searching flights, hotels, etc [17][18] - A compression step is used by sub-agents to synthesize research findings into comprehensive mini-reports before returning to the supervisor, mitigating context window overload [21][23] - The supervisor analyzes findings from sub-agents to either complete research or continue with follow-up questions [25] - Final report generation is done in a single shot using all condensed research findings [5][27] Implementation & Configuration - The agent is built on Langraph and can be run locally by cloning the Open Deep Research repository [29] - Configuration involves setting API keys for models (default OpenAI) and search tools (default Tavily) [30] - Langraph Studio can be used for iterating and testing the agent with different configurations [32] - The agent is highly configurable, allowing users to choose between default or model provider native search tools, connect to MCP servers, and select models for different steps [33][34] Application & Output - The agent can be used for complex research tasks, such as planning a trip, by iteratively calling tools and searching the web [2] - The agent provides a final report with an overview, flight options, transportation options, accommodation options with booking links, a sample itinerary, and a list of sources [36] - Open Agent Platform provides a UI to configure and try out the research agent without cloning the code [37]

Financial Data and Key Metrics Changes - In Q1 2025, the company delivered over 92,000 vehicles, a year-over-year increase of 15.5%, resulting in total revenues of RMB 25.9 billion [7][17] - Total revenues increased by 1.1% year-over-year but decreased by 41.4% quarter-over-quarter [17] - Vehicle sales contributed RMB 24.7 billion, up 1.8% year-over-year but down 22.1% quarter-over-quarter [17] - Gross profit was RMB 5.3 billion, up 0.6% year-over-year but down 40.7% quarter-over-quarter [18] - Net income was RMB 89.1 million, up 9.4% year-over-year but down 81.7% quarter-over-quarter [21] Business Line Data and Key Metrics Changes - The Li L Series continues to lead in the eREV segment, with a market share of 14.1% in April 2025 [8] - The new Li Mega Home model accounted for over 90% of Li Mega orders since its launch, indicating strong user recognition [9][10] - The company plans to increase production capacity to achieve 2,500 to 3,000 monthly deliveries by July 2025 [11] Market Data and Key Metrics Changes - The company has maintained its position as the sales champion among Chinese auto brands for 14 consecutive months [7] - The overall NEV market over RMB 200,000 is expected to reach approximately 3.8 million units in 2025, with 2.1 million being BEVs [48][37] Company Strategy and Development Direction - The company aims to become a top-tier player in the premium BEV market, with a robust pipeline for upcoming models [11] - The launch of the Li Mega Home reflects the company's focus on multi-generational family mobility needs [9] - The company is expanding its sales and servicing network, targeting growth in Tier 4 and Tier 5 cities through the STAR program [16][50] Management's Comments on Operating Environment and Future Outlook - Management expressed confidence in achieving monthly deliveries of the new L series returning to 50,000 units soon [27] - The company expects Q2 2025 deliveries to be between 123,000 to 128,000 units, reflecting a year-over-year increase [11][23] - Management emphasized the importance of user-centric innovation and the integration of AI in future vehicle experiences [70] Other Important Information - The company has built the largest urban highway supercharging network in China, with plans to expand to 4,000 stations by the end of the year [12] - The Li Halo OS has been open-sourced to empower industry partners and foster progress [15] Q&A Session Summary Question: How will Li Auto maintain sales growth against aggressive competition? - Management noted that the new L series has shown healthy growth, with weekly sales exceeding 10,000 units and a market share of 14.7% in the NEV market [27] Question: When will Li Auto consider selling sedans? - Management indicated that after reaching a revenue scale of RMB 300 billion, they will review market conditions and user demand for launching sedan models [29] Question: What is the intelligence level of Li HelloOS and its cost implications? - Management highlighted strong interest in HaloOS from various industry players and emphasized its advantages over traditional operating systems [35] Question: What are the key selling points of the upcoming i8 model? - The i8 will feature innovative styling, high voltage charging technology, and is expected to be a top choice for NEV owners [44] Question: What is the company's target leverage ratio and payable cycle? - Management aims to maintain payable days between two to four months, ensuring healthy relationships with suppliers [41] Question: What are the expectations for the export business? - Management stated that entering overseas markets is part of their long-term strategy, aiming for 30% of overall sales to come from international markets [62]

Core Insights - Meta Platforms is developing LLaMa, an open-source large language model (LLM) aimed at competing with ChatGPT and other AI technologies [1][2] - The current stock forecast for Meta Platforms indicates a potential upside of 18.38%, with a 12-month price target of $719.26 [1][7] Group 1: LLaMa Overview - LLaMa is a large language model trained on extensive human-generated text, enabling it to generate human-like responses [1] - Meta differentiates LLaMa by making it more open-source compared to competitors, allowing customization for various use cases, although some restrictions exist [2][3] Group 2: Financial Implications - Currently, Meta is not directly monetizing LLaMa but may generate revenue through licensing to large enterprises and increased engagement on its platforms [4][5] - The advertising business has seen a 68% rise since the beginning of 2024, partly attributed to the benefits derived from LLaMa [7] Group 3: Future Opportunities - LLaMa is expected to enhance Meta's advertising capabilities, with improved personalization leading to increased advertising revenue over time [8] - There is potential for future monetization of LLaMa, but it is likely to remain open-source, fostering a large ecosystem of applications that could drive ad revenue [9][10] - LLaMa also supports Meta's Reality Labs segment, contributing to innovations like AI-powered Ray-Ban glasses [10][11]