Latent Space: The AI Engineer Podcast
swyx + Alessio
The podcast by and for AI Engineers! In 2024, over 2 million readers and listeners came to Latent Space to hear about news, papers and interviews in Software 3.0.We cover Foundation Models changing every domain in Code Generation, Multimodality, AI Agents, GPU Infra and more, directly from the founders, builders, and thinkers involved in pushing the cutting edge. Striving to give you both the definitive take on the Current Thing down to the first introduction to the tech you'll be using in the next 3 months! We break news and exclusive interviews from OpenAI, Anthropic, Gemini, Meta (Soumith Chintala), Sierra (Bret Taylor), tiny (George Hotz), Databricks/MosaicML (Jon Frankle), Modular (Chris Lattner), Answer.ai (Jeremy Howard), et al.Full show notes always on https://latent.space
Episodes
Mentioned books
58 snips
Dec 23, 2023 • 3h 20min
NeurIPS 2023 Recap — Best Papers
Hosts recap the NeurIPS 2023 conference, discussing best papers and influential topics such as direct preference optimization for language models, scaling data constraint language models, developing a visual intelligent assistant, understanding bunny boxes with GPT4, and using Tool Former to improve language models. They also explore using GPT-4 to play Minecraft, evaluating cognitive capacities through diverse tasks, analyzing language models' performance in planning tasks, and the impact of foundation models on AI systems.
50 snips
Dec 20, 2023 • 59min
The AI-First Graphics Editor - with Suhail Doshi of Playground AI
Suhail Doshi, CEO and co-founder of Playground AI and former Mixpanel chief, dives into the fascinating world of AI-driven graphics editing. He shares insights on the evolution of image generation tools, highlighting real-time preview rendering and style filtering features. The discussion highlights community feedback's critical role in shaping user-friendly tools. Doshi also touches on the ethical implications of AI-generated content and the importance of democratizing digital art for non-designers, unlocking creativity for all.
150 snips
Dec 14, 2023 • 1h 20min
The "Normsky" architecture for AI coding agents — with Beyang Liu + Steve Yegge of SourceGraph
In this engaging discussion, Beyang Liu, co-founder of Sourcegraph, and Steve Yegge, VP of Engineering at Sourcegraph, share their journey from Google to revolutionizing code search. They explore the creation of 'Cody', an AI coding assistant, and the challenges of automating software development. The duo also delves into the historical debate between Chomsky and Norvig, discussing how their innovative 'normsky' architecture enhances coding efficiency. With insights on the future of open-source AI and the importance of context in coding, this masterclass is a treasure trove for tech enthusiasts.
50 snips
Dec 8, 2023 • 1h 4min
The Busy Person's Intro to Finetuning & Open Source AI - Wing Lian, Axolotl
Wing Lian, founder of the Open Access AI Collective and maintainer of Axolotl, shares his journey from web scraping to fine-tuning AI models. He highlights how community efforts and open-source innovations are transforming the AI landscape. Key discussions include navigating hyperparameter tuning, advanced techniques like LoRa, and the ethical considerations of AI licensing. Wing emphasizes the importance of developer tools and community feedback in refining AI applications, all while advocating for collaborative ventures in the space.
70 snips
Nov 29, 2023 • 52min
Notebooks = Chat++ and RAG = RecSys! — with Bryan Bischof of Hex Magic
Bryan Bischof, Head of AI at Hex Magic, shares his extensive experience in data science, previously enhancing strategies at companies like Blue Bottle Coffee and Stitch Fix. He discusses the resurgence of notebooks as a user-friendly interface for AI, dubbing them 'Chat++' for their superior editing capabilities. The conversation also dives into the potential of Retrieval-Augmented Generation (RAG) in creating efficient SQL queries, and the importance of innovative design in AI tools to better serve non-technical users.
49 snips
Nov 17, 2023 • 53min
The State of Silicon and the GPU Poors - with Dylan Patel of SemiAnalysis
Dylan Patel, author of the SemiAnalysis blog, offers insights into the semiconductor industry and GPU dynamics. He discusses the burgeoning divide between the 'GPU rich' and 'GPU poor,' highlighting a significant increase in GPU production forecasted for next year. Patel shares thoughts on AMD's strategic sales and the impact of AI on GPU demand. He also delves into the complexities of GPU utilization for training large models, emphasizing performance optimization and the evolving landscape of semiconductor supply chains.
25 snips
Nov 8, 2023 • 2h 22min
AGI is Being Achieved Incrementally (DevDay Recap - cleaned audio)
In this discussion, Simon Willison, a software developer and author, joins AI researchers and startup founders to reflect on OpenAI's DevDay announcements. They dive into the exciting advancements like GPT-4 Turbo, the risks of prompt injection, and the shift from plugins to GPTs, highlighting the implications for app development. Reid Robinson shares insights on Zapier's AI actions, while Shreya Rajpal emphasizes the importance of guardrails for LLM applications. The conversation also explores new API features and community-driven innovations in AI.
69 snips
Nov 8, 2023 • 2h 23min
AGI is Being Achieved Incrementally (OpenAI DevDay w/ Simon Willison, Alex Volkov, Jim Fan, Raza Habib, Shreya Rajpal, Rahul Ligma, et al)
Joining the conversation are Simon Willison, a software engineer known for his expertise in Python, and Alex Volkov, an AI researcher from Weights & Biases. They delve into the latest from OpenAI’s DevDay, breaking down the new features like GPT-4 Turbo and stateful APIs. Raza Habib discusses the real-world implications of foundation models, while Surya Dantuluri shares insights on evolving chatbot functionalities. Reid Robinson touches on Zapier's AI integration, and the group examines innovative applications and the future of multimodal capabilities in AI.
47 snips
Nov 3, 2023 • 1h 7min
Beating GPT-4 with Open Source LLMs — with Michael Royzen of Phind
In this discussion, Michael Royzen, the founder of Phind, shares his journey from computer vision to creating an innovative search engine for programmers. He discusses Phind's unique approach to answering technical questions and helping users implement solutions. The conversation dives into the evolution of AI in programming, including the impact of GPT-4 and the competition between open-source models. Michael also highlights the importance of user engagement and how his team navigates growth and branding challenges in a fast-paced tech landscape.
18 snips
Oct 26, 2023 • 39min
Powering your Copilot for Data – with Artem Keydunov of Cube.dev
Artem Keydunov, co-founder of Cube.dev, discusses the evolution of natural language understanding in data analytics. He shares insights on transforming StatsBot into Cube, a semantic layer for better data querying. Artem delves into the challenges faced by AI in accurately interpreting SQL queries, emphasizing the risks of hallucinations in AI-generated SQL. He also highlights the necessity of a semantic layer in business intelligence, showcasing how AI tools can become invaluable co-pilots for data-driven decision-making.
