We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
cover of episode Beyond the Model: Taking AI From Prototype to Production with Unstructured’s Brian Raymond

Beyond the Model: Taking AI From Prototype to Production with Unstructured’s Brian Raymond

2024/10/31
logo of podcast Founded & Funded

Founded & Funded
AI Deep Dive AI Chapters Transcript
People
B
Brian Raymond
K
Karan Mehandru
Topics
Brian Raymond: 我认为，将生成式AI工作流程和应用从原型阶段转化为生产阶段非常困难，但正在逐步实现。我们专注于为基础模型提供数据，并帮助用户成功地将其应用于业务用例。AI领域正在从垂直应用转向强大的水平应用，这需要数据预处理和摄取等能力。模型参数规模的增加、数据量的增长以及知识的编码都使得模型能力越来越强大，但延迟、成本和性能等问题依然存在。几乎所有公司都希望拥有一个能够访问所有数据、了解组织所有信息、始终保持最新且运行成本低廉的全知型基础模型能力，实现这一目标是一个漫长的过程。在2022年冬季和2023年，知识图谱在RAG中的地位曾被暂时取代，但后来又重新回归，这凸显了数据摄取和预处理的重要性。Unstructured等公司对于AI的充分发挥至关重要，因为它们负责数据预处理和摄取，这是AI应用的“第一英里”。目前AI领域的投资回报率存在不确定性，但AI的范式转变不可逆转。AI领域的创新速度很快，但目前投资是否合理，以及投资者的预期能否实现，仍存在疑问。模型正在变得越来越大、性能越来越好，同时也越来越小、性能越来越好，这表明AI技术将应用于生活的方方面面。目前评估AI的生产和营收还为时尚早，因为AI技术发展速度很快，而且人们往往高估短期内能取得的成就，低估长期内能取得的成就。Unstructured公司正在适应AI市场的发展变化，并专注于开发者体验、软件基础和与客户的沟通。大型语言模型的快速发展降低了微调的需求，促使公司更多地关注架构和模型的利用方式。Unstructured公司始终专注于开发者体验和软件基础，并通过快速迭代产品和与客户沟通来保持竞争力。我们希望成为连接人类生成数据和基础模型的关键桥梁，并为大型组织提供支持。 Karan Mehandru: 公共部门正在快速采用AI技术，Unstructured公司在帮助公共部门解决AI应用中的数据问题方面发挥着重要作用。关于AI和AI技术栈的一些普遍观点可能并不完全准确。人们对AI的看法存在两种截然相反的观点：一种是认为AI即将实现通用人工智能，另一种是认为AI是一个巨大的泡沫，毫无价值。AI技术的发展离不开众多工程师的协作，他们通过开源和在线社区等方式共同努力，推动了AI技术的发展。

Deep Dive

Chapters
This chapter explores the shift in AI focus from addressing hallucinations to achieving precision at scale. It discusses the challenges and progress in creating omniscient foundation models that are both accurate and cost-effective.
  • Shift from addressing AI hallucinations to achieving precision at scale.
  • Desire for omniscient foundation models that are accurate, up-to-date, and cheap to operate.
  • Challenges in achieving this goal.

Shownotes Transcript

In this episode of Founded & Funded, Madrona Managing Director Karan Mehandru hosts Brian Raymond, founder and CEO of 2024 IA40 winner Unstructured. Unstructured turns raw, messy, unstructured data into something AI tools can actually process — a typically expensive and time-consuming process for companies hoping to save money with the tools themselves.

Brian and Karan explore some of the biggest trends in AI, including moving from prototype to production, the rise and fall of vertical applications, the shift toward powerful horizontal solutions, whether the ROI will be there on all this capital investment into AI, and the trough of disillusionment and why expectations in AI often don't match reality. They also discuss how an unlikely source, the public sector, is driving AI innovation today.

It's a conversation you won't want to miss.

Full Transcript: https://www.madrona.com/prototype-to-production-ai-insights-unstructured-brian-raymond)

Chapters:

(00:00) Introduction  (01:14) Trends in AI: From Vertical to Horizontal Solutions  (01:47) The First Mile of AI: Preprocessing and Data Ingestion  (03:43) Displacement of knowledge graphs for RAG  (06:19) The AI Stack: Layers and Competition  (08:55) Is the current level of investment in AI justified?  (11:19) The Speed of Change in AI(12:47) Public Sector vs. Private Sector Roles in AI Innovation(16:52) "Two completely incongruous narratives"(19:15) Advice for AI Entrepreneurs(21:49) Unstructured's Future