What is Instructure and How It Empowers Developers?

TL;DR

Instructure is a powerful tool designed for developers, enabling them to extract structured data, create knowledge graphs, and enhance query understanding using language models. Its focus on standardization and efficiency allows for improved observability across systems, making it an essential asset in the AI engineering landscape.

Transcript

hey everyone welcome to the laden space podcast this is alesio partner at CTO and residents at deible partners and I'm joined by my co-host swix founder of small AI hello we're back in the remote studio with Jason Lou from instructor welcome Jason hey there thanks for having me Jason you are extremely famous so I don't know what I'm going to do int... Read More

Key Insights

• 📈 Instructure is a powerful tool for developers, offering capabilities such as structured data extraction, knowledge graph creation, and query understanding.
• 🎰 The tool addresses the need for standardization in machine learning systems, providing a unified framework for improved observability and efficiency.
• 😒 Instructure's similarity search system uses CLIP and GPT-3 embeddings to enhance recommendation systems.
• 🧡 Language models, such as GPT-3, have evolved over time, making them more viable for a wide range of applications.
• 👻 Instructure's focus on workflows and data structures allows for better experimentation and optimization.
• 💄 The tool's SDK simplifies the integration of language models, making it accessible to developers.
• 🎮 Jason Lou's decision to focus on consulting rather than seeking venture capital funding reflects his desire for optionality and control over the direction of Instructure.
• 🥹 Agency and courage are essential for pursuing entrepreneurial endeavors, even if the expected financial gains might not be as high as in venture-backed startups.
• 👨‍💻 Prompts should be code-like objects to maintain control and flexibility in using language models effectively.
• ❎ Experimentation and failure are valuable, even if the outcomes are negative, as they provide insights and potential paths for future improvement.

Questions & Answers

Q: How did Jason Lou's perception of language models change over time?

Jason initially had doubts about the practicality of language models but later realized the value they bring to various applications, prompting him to transition from computer vision to language models.

Q: What is the primary purpose of Instructure?

Instructure is a tool that enables developers to extract structured data, build knowledge graphs, and enhance query understanding using language models.

Q: How did Jason Lou address the need for standardization at Stitch Fix?

Jason developed Instructure to standardize the machine learning systems at Stitch Fix, enabling different teams to use a unified framework and improve observability across the entire pipeline.

Q: How does Instructure's similarity search system work?

Instructure leverages CLIP, GPT-3 embeddings, and an FE model to create image embeddings that capture descriptions and sales data. These embeddings are then used to enhance the recommendation system.

Summary & Key Takeaways

• Jason Lou, founder of Instructure, discusses his journey from being skeptical of language models to becoming a user and advocate of them at Stitch Fix and beyond.

• Instructure is a versatile tool that allows developers to extract structured data, build graphs, and improve query understanding using language models.

• The tool's key applications include extracting structured data, creating knowledge graphs, and enhancing query understanding through defined schemas.