Stories with Tag Lago

• Why DeepSeek had to be open source

  permalink

  Posted: 2025-01-29 15:37:31

  Verbose tl;dr

  DeepSeek, a platform offering encoder APIs for developers, chose to open-source its core technology due to the inherent difficulty in building trust with users regarding data privacy and security when handling sensitive information like codebases and internal documentation. By open-sourcing, DeepSeek aims to foster transparency and allow users to self-host, ensuring complete control over their data. This approach mitigates concerns around vendor lock-in and allows the community to contribute to the project's development and security, ultimately building greater trust and fostering wider adoption.

  The blog post "Why DeepSeek had to be open source," published by Lago, details the strategic rationale behind DeepSeek's decision to embrace an open-source model for their encoder technology. DeepSeek, a company specializing in AI-powered code search, faced the formidable challenge of establishing trust and widespread adoption within the developer community, a group known for its preference for open and transparent tools. The closed-source approach presented a significant obstacle to achieving this goal, as developers are often hesitant to entrust proprietary systems with access to their valuable and often sensitive codebases.

  The blog post articulates that open-sourcing the DeepSeek encoder allows developers to thoroughly inspect and understand the underlying mechanisms of the code search technology, fostering trust and confidence in its operation. This transparency eliminates the "black box" effect inherent in closed-source solutions, allowing developers to verify the encoder's security, efficiency, and accuracy firsthand. By providing full visibility into the code, DeepSeek empowers the community to actively contribute to the project, identifying potential vulnerabilities or areas for improvement, leading to a more robust and reliable system. This collaborative development model also benefits DeepSeek directly by leveraging the collective expertise of the open-source community, accelerating the pace of innovation and refinement.

  Furthermore, the open-source approach directly addresses the critical issue of data privacy, a major concern for developers when utilizing third-party code analysis tools. By making the encoder's source code publicly available, DeepSeek demonstrates a commitment to transparency and allows developers to verify that the encoder does not exfiltrate sensitive data or intellectual property. This reassurance is essential for gaining the trust of organizations and individual developers, paving the way for wider adoption of the technology.

  The post also emphasizes the strategic advantage of open-sourcing the encoder while maintaining the proprietary nature of the vector database technology. This approach allows DeepSeek to offer a commercially viable product while simultaneously benefiting from the open-source community's contributions to the encoder. This dual approach strikes a balance between fostering community engagement and ensuring the long-term sustainability of the business.

  Finally, the blog post positions the open-sourcing of the DeepSeek encoder as a crucial step in establishing a robust ecosystem around their technology. By encouraging community involvement and contributions, DeepSeek aims to cultivate a vibrant and active developer ecosystem, driving further innovation and accelerating the adoption of AI-powered code search tools. The open-source model is presented as a catalyst for growth and collaboration, laying the foundation for a thriving community that benefits both developers and DeepSeek.

  • Open Source
  • deepseek
  • Search Engine
  • Code Search
  • Lago
  • developer tools
  • Software Development
  • Search Technology
  • OSS

  Summary of Comments ( 242 )
  https://news.ycombinator.com/item?id=42866201

  Verbose tl;dr

  Hacker News users discussed the open-sourcing of DeepSeek, primarily focusing on the challenges of monetizing open-source AI infrastructure. Many commenters were skeptical of Lago's business model, questioning how they could successfully build a proprietary offering on top of an open-source core, especially given the intense competition in the vector database space. Some suggested that open-sourcing DeepSeek was a necessary move due to the difficulty of attracting paying customers for a closed-source product. Others pointed out potential advantages, such as faster iteration and community contributions, but remained unconvinced of long-term viability. Several users expressed a desire for more technical details about DeepSeek's implementation and performance compared to existing solutions. The most compelling comments revolved around the inherent tension between open-sourcing and profitability in the current AI landscape.

  The Hacker News post "Why DeepSeek had to be open source" (linking to a blog post about the open-sourcing of a vector database called DeepSeek) generated a moderate amount of discussion, with several commenters focusing on the challenges and tradeoffs inherent in open-sourcing complex infrastructure software.

  One compelling line of discussion revolved around the difficulty of monetizing open-source infrastructure projects. A commenter pointed out the "challenging economics" of open-sourcing core infrastructure, noting that "it's hard to build a business on top of open core, especially for infrastructure software" and suggested that open-sourcing could be a last resort due to difficulties in acquiring customers. This spurred further discussion about the potential downsides of "open-core" business models, with some expressing skepticism about their long-term viability.

  Another commenter highlighted the specific complexities of vector databases, stating that they are "notoriously hard to operate" and require significant expertise. This raises the question of whether open-sourcing DeepSeek might actually hinder its adoption due to the increased burden on users to manage and maintain the database themselves. They further suggested that a managed service offering would likely be more appealing to many potential users, echoing the sentiment about the difficulties of the open-core model in this space.

  Several comments touched upon the competitive landscape of vector databases, mentioning alternatives like Pinecone, Weaviate, and Qdrant. One commenter expressed surprise that DeepSeek hadn't already been acquired, suggesting that the vector database space is attracting significant interest and investment.

  Finally, a few commenters questioned the blog post's premise that DeepSeek "had to be" open-sourced, suggesting that this framing might be a marketing tactic rather than a genuine necessity. They proposed alternative explanations, such as the possibility that the company was struggling to attract paying customers or that open-sourcing was a way to gain community contributions and improve the software's quality.

  In summary, the comments on Hacker News primarily focused on the business implications of open-sourcing DeepSeek, the technical challenges of running vector databases, and the competitive dynamics of the market. Several commenters expressed skepticism about the viability of open-sourcing complex infrastructure software and suggested that a managed service might be a more successful approach.

• Solving complex billable metrics with custom SQL expressions in Lago

  permalink

  Posted: 2025-01-27 12:12:47

  Verbose tl;dr

  Lago's blog post details how their billing platform now supports custom SQL expressions for defining billable metrics. This allows businesses with complex pricing models greater flexibility and control over how they charge customers. Instead of relying on predefined metrics, users can now write SQL queries directly within Lago to calculate charges based on virtually any data they collect, including custom events and attributes. This simplifies the implementation of usage-based billing scenarios like charging per API call with specific parameters, tiered pricing based on aggregate usage, or dynamic pricing based on real-time data. The post emphasizes how this feature reduces development time and empowers product and finance teams to manage billing logic without extensive engineering involvement.

  The Lago blog post, "Solving complex billable metrics with custom SQL expressions in Lago," details how Lago's platform now allows users to define highly customized billable metrics using SQL expressions, offering greater flexibility and control over billing logic. Traditionally, subscription billing systems struggle with complex, usage-based pricing models. Lago addresses this challenge by enabling users to leverage the power and expressiveness of SQL directly within their billing engine. This allows for the creation of intricate metrics tailored to unique business requirements, moving beyond simple, pre-defined metrics.

  The post emphasizes the limitations of traditional subscription management platforms, where metrics are often rigid and lack the granularity needed for complex scenarios. For instance, if a business wants to charge based on a specific interaction or a combination of factors, traditional systems may fall short. Lago's custom SQL expressions provide a solution by allowing users to define billable metrics based on any data stored within their Lago instance. This empowers businesses to implement sophisticated pricing models, such as tiered pricing based on specific usage patterns, or hybrid models combining usage with subscription fees.

  The blog post provides a practical example of calculating the number of weekly active users (WAU) with a custom SQL expression, demonstrating how this feature can be used in a real-world scenario. This example highlights the flexibility and power of the SQL-based approach, allowing businesses to calculate metrics that are precisely aligned with their specific definition of an "active user." This granular control enables more accurate and transparent billing, reducing the risk of disputes and improving customer relationships.

  Furthermore, the post emphasizes the extensibility of this feature, suggesting that any aggregatable data within the Lago platform can be used to construct custom billable metrics. This opens up numerous possibilities for innovative pricing models and allows businesses to tailor their billing to reflect the true value delivered to their customers. By bringing the power of SQL to billing metric definition, Lago simplifies the implementation of complex pricing structures, enabling businesses to experiment with and adapt to evolving market demands without being constrained by rigid billing systems. This ultimately allows businesses to focus on their core product and value proposition rather than wrestling with intricate billing logic.

  • SaaS
  • Billing
  • Metrics
  • SQL
  • Custom SQL
  • Lago
  • Billable Metrics
  • Subscription Billing
  • Usage-Based Billing
  • Pricing
  • Fintech
  • software engineering
  • Software Development
  • API

  Summary of Comments ( 0 )
  https://news.ycombinator.com/item?id=42840303

  Verbose tl;dr

  Hacker News users discuss Lago's approach to flexible billing using custom SQL expressions. Some express concerns about the potential complexity and debugging challenges of using SQL for this purpose, suggesting simpler alternatives like formula-based systems. Others highlight the power and flexibility SQL offers for handling complex billing scenarios, especially for businesses with intricate pricing models. A few commenters question the performance implications of using SQL queries for real-time billing calculations and suggest pre-aggregation or caching strategies. There's also discussion around the trade-off between flexibility and auditability, with concerns about the potential difficulty in understanding and verifying SQL-based billing logic. Some users share their experiences with similar systems, emphasizing the importance of thorough testing and validation.

  The Hacker News post "Solving complex billable metrics with custom SQL expressions in Lago" at https://news.ycombinator.com/item?id=42840303 has generated several comments discussing the merits and drawbacks of Lago's approach to billing using custom SQL expressions.

  One commenter expresses concern about vendor lock-in, suggesting that relying on a specific vendor's SQL dialect for defining billing logic could create difficulties if migrating to a different platform in the future. They propose that a standardized approach, perhaps using something like CEL (Common Expression Language), might be a better long-term strategy.

  Another commenter points out the inherent complexity of billing systems and argues that SQL, despite its potential for vendor lock-in, is a reasonable choice due to its widespread familiarity and the existing tooling available for working with it. They acknowledge that no single solution will be perfect for every scenario but suggest that SQL offers a good balance between flexibility and accessibility. This comment sparked further discussion about the benefits of standardization versus the practicality of using existing, well-understood tools.

  Building on the vendor lock-in concern, another user notes the potential for "gotchas" within custom SQL implementations. They highlight that subtle differences in how SQL dialects handle specific functions or data types could lead to unexpected billing discrepancies. This reinforces the argument for careful consideration and thorough testing when employing custom SQL for billing.

  A different perspective is offered by a commenter who appreciates the transparency and control that custom SQL expressions can provide. They argue that being able to directly define billing logic in SQL allows for greater flexibility and customization compared to relying on pre-defined billing models. This, they suggest, can be particularly beneficial for businesses with unique or complex billing requirements.

  There's also a brief discussion about the potential performance implications of using custom SQL for billing. One commenter raises the question of how Lago handles the execution of these SQL expressions and whether it could introduce performance bottlenecks, especially with large datasets. This concern, however, wasn't addressed directly in the comments.

  Finally, some commenters mention alternative approaches to billing, including using tools like Stripe Billing or building custom in-house solutions. These suggestions highlight the range of options available to businesses and emphasize the importance of choosing the right solution based on specific needs and constraints.