Infinite Curiosity Pod with Prateek Joshi

Matt Welsh is the cofounder and CEO of Fixie, an automation platform for LLMs. It allows developers to build natural language agents that connect to your data, talk to APIs, and solve complex problems. They've raised $17M from investors such as Redpoint, Madrona, Zetta, SignalFire, Bloomberg Beta, and more. He has previously held roles at OctoML, Apple, Xnor, and Google. He was a Professor of Computer Science at Harvard and has a PhD in Computer Science from UC Berkeley. In this episode, we cover a range of topics including: - LLMs as the new computational engine - What can LLMs do well and where are the gaps - Fine-tuning vs In-context learning - Smart Agents - Few shot learning - Use cases of Fixie Matt's favorite book: The Amazing Adventures of Kavalier & Clay (Author: Michael Chabon)--------Where to find Prateek Joshi:  Newsletter: https://prateekjoshi.substack.com Website: https://prateekj.com LinkedIn: https://www.linkedin.com/in/prateek-joshi-91047b19 Twitter: https://twitter.com/prateekvjoshi 

Aditya Naganath is an investor at Kleiner Perkins, a legendary VC firm with investments in companies such as Google, Amazon, Spotify, Doordash, Slack, Intuit, UiPath, Figma, and many more. Prior to this, he was a product manager at Google focusing on growth initiatives for the SMB monetization team within emerging markets. He has previously been in technical roles at Palantir, Twitter, and Nextdoor. He earned a patent during his time at Twitter for a technical analytics product he co-created. And he has a bachelor’s degree in Computer Science from Columbia University and an MBA from Stanford. In this episode, we cover a range of topics including: - Uses of AutoGPT - How to intervene when AutoGPT gets stuck and how to provide feedback - ChatGPT Plugins - Generative AI in the enterprise - Context windows of LLMs - What new emergent properties can be uncovered if LLMs can reason over larger context windows - LLMs for developers - AI landscape and investment opportunities Aditya's favorite book: Factfulness (Author: Hans Rosling)--------Where to find Prateek Joshi:  Newsletter: https://prateekjoshi.substack.com Website: https://prateekj.com LinkedIn: https://www.linkedin.com/in/prateek-joshi-91047b19 Twitter: https://twitter.com/prateekvjoshi 

Elad Gil is an entrepreneur, operator, and and investor. He has invested in and advised companies such as AirBnB, Stripe, Figma, Airtable, Samsara, Coinbase, Flexport, Gitlab, Gusto, Instacart, Notion, Pinterest, Rippling, and many more. He was the cofounder and Chairman at Color Genomics. Prior to that, was the VP of Corporate Strategy at Twitter. He was the cofouder and CEO MixerLabs, which got acquired by Twitter. He has held roles at Google and McKinsey. He has a PhD from MIT. In this episode, we cover a range of topics including:  - How will the AI market structure evolve? - What products can startups build to make LLMs faster/better/cheaper? - How is Generative AI being used in the enterprise today? - How are customers spending money for AI-related work? - What is AutoGPT and how can businesses use it? - What AI copilot products can be built? - How are LLMs being used for software development? - How will the AI infrastructure market shape up? - Building a moat for AI-infused businesses - Open source opportunities in AI - What's the state of play in the AI application market? -------- Where to find Prateek Joshi: Newsletter: https://prateekjoshi.substack.com Website: https://prateekj.com LinkedIn: https://www.linkedin.com/in/prateek-joshi-91047b19 Twitter: https://twitter.com/prateekvjoshi 

Jerry Liu is the cofounder and CEO of LlamaIndex. He is the creator of the open source tool that's also named LlamaIndex, which provides a central interface to connect your LLMs with external data. He has previously held roles at Quora, Uber, and Robust Intelligence. He has a computer science degree from Princeton. In this episode, we cover a range of topics including:  - What can LLMs do well and where are the gaps - How to connect LLMs to external data - What does LlamaIndex do - Fine-tuning LLMs - In-context learning - Application that are getting built on top of LlamaIndex Jerry's favorite books: Harry Potter series (Author: J.K. Rowling) and The Hard Thing About Hard Things (Author: Ben Horowitz)--------Where to find Prateek Joshi: Newsletter: https://prateekjoshi.substack.com Website: https://prateekj.com LinkedIn: https://www.linkedin.com/in/prateek-joshi-91047b19 Twitter: https://twitter.com/prateekvjoshi

Matt Turck is a Partner at FirstMark Capital where he focuses primarily on early-stage enterprise investing. He is particularly active in the data, machine learning, and AI. Since 2011, he has been organizing Data Driven NYC, the largest data/AI community in the US. Since 2012, he has been publishing an annual landscape of the data/AI industry, the MAD report. Earlier in his career, he was the co-founder of TripleHop, an enterprise search software startup that was acquired by Oracle. Immediately prior to FirstMark, he was a Managing Director at Bloomberg Ventures, the incubation arm of Bloomberg LP, which he helped start.In this episode, we cover a range of topics including: - AI going mainstream with ChatGPT- ChatGPT Plugins- Generative AI in the enterprise- AI startup landscape- "Copilot for X" products- What gets him excited about an investment opportunity- The moat of AI-infused businesses- GTM motion of successful AI startupsMatt's favorite books: All the books in the Incerto series (Author: Nassim Nicholas Taleb)--------Where to find Prateek Joshi: Newsletter: https://prateekjoshi.substack.com Website: https://prateekj.com LinkedIn: https://www.linkedin.com/in/prateek-joshi-91047b19 Twitter: https://twitter.com/prateekvjoshi 

Bob van Luijt is the cofounder and CEO of Weaviate. They build, maintain, and commercialize the open-source vector database called Weaviate. They've raised funding from amazing investors such as Index and NEA. He is the Chairman of Creative Software Foundation, a nonprofit located in The Netherlands with a sole focus on open source and creative software. In this episode, we cover a range of topics including: - What are vector embeddings - What's a good vector representation - What is vector search and why do we need it - What is a vector database and why do we need it - How does a vector database work in practice - Applications that use vector databases Bob's favorite book: The Creative Act (Author: Rick Rubin)--------Where to find Prateek Joshi:  Newsletter: https://prateekjoshi.substack.com Website: https://prateekj.com LinkedIn: https://www.linkedin.com/in/prateek-joshi-91047b19 Twitter: https://twitter.com/prateekvjoshi 

Jake Saper is a General Partner at Emergence Capital, one of the best SaaS VCs in the world. They've invested in iconic companies such as Salesforce, Zoom, Veeva, Gusto, Box, and Bill.com. He was previously at Kleiner Perkins as part of their Green Growth Fund and has an MBA from Stanford. In this episode, we cover a range of topics including: - Limitations of generative AI in B2B SaaS - Iterative AI vs Generative AI - How to solve the user engagement problem with AI - Coaching networks - Pricing model for AI companies - How to build a moat - How you invest in startups - GTM motion of successful AI companies - AI market trends Jake's favorite book: Freedom at Midnight (Authors: Larry Collins and Dominique Lapierre) --------Where to find Prateek Joshi: Newsletter: https://prateekjoshi.substack.com Website: https://prateekj.com LinkedIn: https://www.linkedin.com/in/prateek-joshi-91047b19 Twitter: https://twitter.com/prateekvjoshi 

Jessica Leao is a Principal at Maverick Ventures, a global investment firm that has been investing in early stage companies for over 25 years. She was previously the founder and GP at The 21 Fund, a $2M seed fund building and investing in the next generation of startups out of Stanford Graduate School of Business. She's been the Managing Partner at the Dorm Room Fund, which is backed by First Round Capital. She's also held roles at Palantir, Fortress Investment Group, and Goldman Sachs. In this episode, we cover a range of topics including: - AI compute landscape - Trends and opportunities in AI infrastructure - "Copilot for X" business model - GTM motion of AI businesses - Using AI in legacy industries - How she invests in startups Jessica's favorite book:  Klara and the Sun (Author: Kazuo Ishiguro) --------Where to find Prateek Joshi:Newsletter: https://prateekjoshi.substack.com Website: https://prateekj.com LinkedIn: https://www.linkedin.com/in/prateek-joshi-91047b19 Twitter: https://twitter.com/prateekvjoshi 

Leo Polovets is a cofounder and General Partner at Susa Ventures, a seed stage firm that has invested in iconic companies such as Robinhood and Flexport. Prior to Susa, he was the second engineering hire at Linkedin where he helped build the first versions of products like LinkedIn Jobs and LinkedIn Groups. He then worked on payment fraud detection algorithms at Google, and was also an early engineer at Factual, where he built data cleaning, processing, and deduping software. And he has an engineering degree from Caltech.  In this episode, we cover a range of topics including: - How to derisk a startup - PMF risk, market risk, and funding risk - How should enterprise SaaS founders think about the first 10 customers (choice of customers, how to find them, how to convince them)  - How should VCs think about providing feedback on pitches (why is it important and how to do it well) - Pricing mechanics - How should seed stage companies think about moat - AI-infused software applications - What gets him excited about an investment opportunity - GTM motion of successful seed stage startups Leo's favorite books: - A Guide to the Good Life (Author: William Irvine) - Traction: How Any Startup Can Achieve Explosive Customer Growth (Authors: Gabriel Weinberg and Justin Mares) --------Where to find Prateek Joshi: Newsletter: https://prateekjoshi.substack.com Website: https://prateekj.com LinkedIn: https://www.linkedin.com/in/prateek-joshi-91047b19 Twitter: https://twitter.com/prateekvjoshi 

Brian Raymond is the founder and CEO of Unstructured. They provide open source libraries and APIs to build custom preprocessing pipelines for labeling, training, and production machine learning pipelines. They've raised a seed round from BCV, Mango, and Essence. He was previously the VP of Global Public Sector at Primer. Prior to that, he has held roles at CIA and the National Security Council. In this episode, we cover a range of topics including: - The need for data preprocessing - LLM use cases that are exciting - What AI use cases are useful to government agencies - How should startups think about building for the government - Building the user base for Unstructured - 4 pillars of the emerging tech stack for LLMs - Open source software Brian's favorite books: Chip War, The Main Enemy, Damascus Station, The Red Sparrow Trilogy--------Where to find Prateek Joshi:Newsletter: https://prateekjoshi.substack.com Website: http://prateekj.com LinkedIn: https://www.linkedin.com/in/prateek-joshi-91047b19 Twitter: https://twitter.com/prateekvjoshi