In this episode, we sat down with Bob van Luijt (https://twitter.com/bobvanluijt), the CEO of Weaviate, diving into the cutting-edge world of vector databases and their role in the AI revolution.Weaviate is an open source, AI-native vector database that helps developers create intuitive and reliable AI-powered applications. Weaviate sets itself apart with its vector search engine that integrates machine learning directly into its core, enabling more nuanced and context-aware search capabilities for AI-driven applications.This conversation explores vector databases (the core infrastructure behind generative models), the role of Retrieval-Augmented Generation (RAG), and how open source is driving commercial use cases.WEAVIATEWebsite - https://weaviate.ioTwitter - https://twitter.com/weaviate_ioBob van Luijt (Co-Founder & Co-CEO):LinkedIn - https://www.linkedin.com/in/bobvanluijtTwitter - https://twitter.com/bobvanluijtMatt Turck:LinkedIn - https://www.linkedin.com/in/turck/Twitter - https://twitter.com/mattturckDATA DRIVEN NYCThis episode of the MAD Podcast was recorded live at Data Driven NYC, an event series organized by FirstMark Capital. The events are free and held monthly in New York, currently with the support of Foursquare.If you wish to attend and be notified of future events, please follow FirstMark on Eventbrite at https://www.eventbrite.com/o/firstmark-capital-221557018301:00 What is RAG?06:20 Why is embedding models is such a hot topic right now?08:06 What is your assessment of RAG?09:53 Generative feedback loops11:46 What is Hybrid Search?15:15 What makes Weaviate special?16:53 What about security?17:45 Does RAG accelerated the need for real-time data?19:27 How to define good vector database? 22:11 What do you think about general purpose databases entering the field of vector-based databases?23:47 Interesting use cases of Weaviate25:27 What’s your sense of the current state of the market?26:53 Open source vs commercial product on Weaviate29:23 How did it all get started?

Last week, we sat down with Alex Rinke (https://twitter.com/alexanderrinke), Co-founder & Co-CEO of Celonis, to explore how AI and automation are transforming business operations at large enterprises. Celonis is the pioneer of "process mining" - the technology that uses graph databases, AI, and automation to analyze processes, find inefficiencies and their root causes, and solve them.Most recently valued at $13B, Celonis is one of the most valuable startups globally. But Alexander and his two co-founders started Celonis while still in college on a $15,000 budget. In this conversation, we talked about the early days of Celonis, how Alex acquired his first enterprise clients without inside industry connections, how Celonis navigates go-to-market for a product with an expansive scope, and much more.CELONISWebsite - https://www.celonis.comTwitter - https://twitter.com/CelonisAlex Rinke (Co-Founder & Co-CEO):Twitter: https://twitter.com/alexanderrinkeLinkedIn: https://www.linkedin.com/in/alexander-rinke-10733061/DATA DRIVEN NYCThis episode of the MAD Podcast was recorded live at Data Driven NYC, an event series organized by FirstMark Capital. The events are free and held monthly in New York, currently with the support of Foursquare. If you wish to attend and be notified of future events, please follow FirstMark on Eventbrite at https://www.eventbrite.com/o/firstmark-capital-221557018300:00 - Intro02:02 - What is Process Mining?05:20 - How Celonis got started07:42 - “We had our first prototype in three weeks”09:36 - Pivotal partnership with ACP12:12 - How did Celonis find product-market-people fit?14:14 - Penetrating the global market16:19 - Technical deep dive into the Celonis’ product19:29 - Celonis finds process gaps completely automatically21:15 - Who is the average user of Celonis inside companies?22:11 - How Celonis uses Generative AI 24:54 - Acquisition of Symbio25:56 - How to keep the fire of innovation inside the team?27:49 - How to bring a very horizontal product to market?32:24 - Scaling yourself as a leader34:15 - Glimpse into the future of Celonis35:37 - Outro

We are so excited today to be joined by Brandon Duderstadt, CEO + Cofounder, and Zach Nussbaum, Machine Learning Engineer, from Nomic AI. They discuss how Nomic AI is building tools like Atlas + GPT4all that enable everyone to interact with AI scale datasets and run models on consumer computers - and - stay tuned for an exciting announcement about their newest product release later in the podcast.Thanks for joining us for the first episode of Season 2 of the MAD Podcast. We will be back to our regular weekly schedule with new conversations with leaders in the Machine Learning, AI and data landscape. If you like this show, you can find the video recording of this episode -- along with many, many more -- on the Data Driven NYC channel on YouTube.NOMIC AIwww.nomic.aitwitter.com/nomic_aiwww.linkedin.com/in/bstadt/www.linkedin.com/in/zach-nussbaum/FIRSTMARKfirstmark.comtwitter.com/FirstMarkCapMatt Turck (Managing Director)www.linkedin.com/in/turck/twitter.com/mattturckData Driven NYC YouTube ChannelFirstMark Capital Eventbrite0:46 - What is Nomic AI & how it got started5:57 - Building GPT4ALL7:23 - Running LLMs on a personal computer16:00 - Nomic Atlas21:33 - Launching Nomic Embed28:10 The Importance of Data in AI31:10 - Benchmarking LLMs32:56 - The Future of Nomic AI36: 22 - Building an AI Startup in New York39:10 - Nomic AI is hiring

Today, we’re thrilled to be joined by Eiso Kant, CTO + Co-Founder of Poolside, the buzzy new AI tool for software development. Eiso and Matt talk about Poolside’s foundational model, the critical role of data quality in AI, the importance of controlling all levels of the stack and the merits of building a global AI company out of Europe, and more. Thank you to everyone who has joined us for Season 1 of the MAD Podcast. We will be taking a short break for the winter holidays and will be back with an exciting new lineup of great speakers for Season 2 on Wednesdays in January. If you like this show, you can find the video recording of this episode -- along with many more -- on the Data Driven NYC channel on YouTube. Important links are in the show notes below. Data Driven NYC YouTube ChannelFirstMark Capital Eventbritetwitter.com/eisokantpoolside.aitwitter.com/mattturcklinktr.ee/mattturckShow Notes: [00:38:00] Introducing Eiso Kant, Co-founder and CTO of the AI startup, Poolside;[00:39:16] Eiso's Background; his journey, from starting as a young programmer to founding several companies, including Source{d}, a pioneer in applying deep learning to software source code;[00:40:33] Formation of Poolside; the collaboration between Eiso and his co-founder, Jason Warner, who was previously the CTO of GitHub and VC with Redpoint Ventures;[00:42:14] Poolside's Vision and potential to improve software development;[00:47:17] Narrowing Vision to Product Development; the importance of sequence in a company's growth, focusing on AI pair programming assistants as a start, moving towards a more autonomous future;[00:50:32] Initial Product Focus, user base, and approach to providing a vertically integrated AI stack for developers;[00:53:05] Reinforcement Learning from Code Execution Feedback;[01:02:29] Data Handling and Synthetic Data Generation; the importance of data quality and Poolside's strategy for generating and refining training data;[01:12:05] Engineering Behind Poolside's AI; the challenges and strategies Poolside is adopting, including building a team of strong engineers and creating a scalable architecture from scratch;[01:16:52] Choosing Europe as a Base for Poolside;[01:20:22] Poolside's Future Plans; the roadmap for Poolside, including launching products and APIs, exploring enterprise solutions, and creating a sustainable revenue-generating business;

Today, we’re joined by Gustavo Sapoznik, Founder and CEO of ASAPP, the generative AI platform transforming contact centers. Matt + Gustavo discuss the magnitude of challenges to overcome in this market, how their AI tech is designed to help humans, the reason smart people should choose working at a startup over Big Tech, and more. This session was recorded live at a recent Data Driven NYC, our in-person, monthly event series. If you are ever in New York, you can find us on Eventbrite by searching for "FirstMark Capital". Events run monthly and are free and open to everyone. And as always, if you enjoy the MAD podcast, please subscribe and feel free to leave us a comment or rating.Data Driven NYC YouTube ChannelFirstMark Capital Eventbriteasapp.comtwitter.com/asapptwitter.com/mattturcklinktr.ee/mattturckShow Notes: [00:00:45] Introducing Gustavo Sapoznik, Founder & CEO of ASAPP, a unicorn AI startup based in New York;[00:01:00] How ASAPP started with a mission to “end bad customer service” after a frustrating phone call Mr. Sapoznik had with his cable provider;[00:02:44] ASAPP’s product philosophy and how the customer service is a three-legged stool with companies, customers, and agents;[00:05:11] How ASAPP automates what they can and augments the rest to make agents more productive;[00:07:12] The evolution of ASAPP’s offerings including how ASAPP technology makes agents more productive;[00:9:16] How ASAPP’s technology reduces response times and improves quality for agents by including transcription, auto complete, and real-time scoring of interactions for quality assurance;[00:13:49] How ASAPP has evolved since 2014; their research-first approach, building in-house AI capabilities, training their own models, and their recent exploration of using open-source checkpoints;[00:15:05] How Mr. Sapoznik hired the guy who ran all NLP research at Google;[00:16:04] How cost, latency, and accuracy in their AI models differentiate ASAPP from common AI APIs available today;[00:18:49] Agent models v. Language models and how ASAPP AI is modularized for large teams with established tech stacks;[00:20:09] Mr. Sapoznik shares insights on selling to large enterprises and why he believes building a sales machine is equally, if not more important, than the product itself;[00:23:08] How to recruit and retain top AI talent;[00:27:42] Lessons learned from working with notable board members, including the three key dimensions of support from a good board: being a sounding board, providing tactical advice and connections, and instilling a sense of accountability and motivation;

Today, we’re excited to chat with Scott Belsky - author, entrepreneur, investor and Chief Strategy Officer at Adobe. Matt + Scott discuss the impact of AI on creative work, how Adobe is incorporating AI across their products, and what the future creative tools landscape might look like.This session was recorded live at a recent Data Driven NYC, our in-person, monthly event series. If you are ever in New York, you can find us on Eventbrite by searching for "FirstMark Capital". Events run monthly and are free and open to everyone. And as always, if you enjoy the MAD podcast, please subscribe and leave us a comment.Data Driven NYC YouTube ChannelFirstMark Capital Eventbritetwitter.com/scottbelskyImplications, by Scott Belskytwitter.com/mattturcklinktr.ee/mattturckShow Notes: [00:53] How Adobe uses AI to enhance user experience, streamline onboarding and automate tasks across their product suite;[01:30] How AI impacts Adobe's business, making creative processes accessible with features like the context bar in Photoshop;[02:13] Firefly's journey: internal decisions, training challenges, and a commitment to using licensed material for ethical AI;[03:58] Moral considerations in Firefly's development: the decision to use licensed material, commercial viability, and addressing user comparisons;[05:52] Adobe's homegrown approach to generative AI models: in-house development and partnerships for specific capabilities like LLM;[06:08] Adobe Sensei's 10-year evolution: developing AI technologies, the non-profit Content Authenticity Initiative, and content credentials establishing asset provenance;[09:17] Adobe's new AI advancements: Firefly Image Model 2, Generative Match, and the vector model for illustration;[11:16] Firefly Editor's revolutionary image editing: dynamically generating pixels, real-time object manipulation, and Adobe's commitment to pushing technological boundaries;[12:41] Rapid integration of AI features: Firefly models and playground, surfacing on a website for user testing, and collaboration within Adobe's design organization;[14:32] How Adobe's AI and data teams are structured and leveraging in-house development for competitive advantage;[15:47] Future of work and creativity: AI's impact on raising the bar for digital experiences, accelerating creative processes, and the evolving landscape of personalized social content;[19:11] Leveraging technology to reduce friction, streamline processes, and unlock creative flow;[20:09] Impact of AI on business models: questioning time-based pricing, anticipating a shift to value-based models, and reconsidering compensation for creative professionals;[21:10] Parallels with historical Internet Service Providers, the rapid evolution of ideas, and reflections on sustainable business models;[24:53] Scott’s criteria for evaluating AI investments: valuing skeptical entrepreneurs, acknowledging temporary uniqueness, and emphasizing empathy with customers;[26:40] Navigating challenges in 2023: Tough decisions for entrepreneurs, evaluating conviction, and the importance of sticking together through the "messy middle”;

Today, we’re joined by Howard Katzenberg, CEO of Glean AI, a machine learning powered accounts payable platform. Matt + Howard discuss Glean’s founding story, how Glean helps CFOs make insight driven choices, and more. This session was recorded live at a recent Data Driven NYC, our in-person, monthly event series. If you are ever in New York, you can find us on Eventbrite by searching for "FirstMark Capital". Events run monthly and are free and open to everyone. And as always, if you enjoy the MAD podcast, please subscribe and leave us a comment. Data Driven NYC YouTube ChannelFirstMark Capital Eventbritetwitter.com/mattturck linktr.ee/mattturckShownotes: [00:00:35] Howard's background;[00:01:15] Challenges with manual FP&A;[00:02:54] Approval Process gap realization and opportunity for Glean AI;[00:04:40] How Glean AI is like “bill.com with a brain”;[00:05:06] Enhanced functionalities beyond basic AP automation;[00:06:32] Glean AI’s Inception and AI Models;[00:07:54] Why Glean AI is unique;[00:08:25] The evolution of Glean AI’s ML stack;[00:10:44] Defensibility and how Glean AI offers vendor pricing insights to its network;[00:12:23] Success stories and customer value;[00:14:47] Future plans for Glean AI;[00:16:39] Navigating industry and technical expertise;[00:18:41] Audience Q&A

Today, we have the pleasure of chatting with Raza Habib, CEO of Humanloop, the platform for LLM collaboration and evaluation. Matt and Raza cover how to understand and optimize model performance, lessons learned about model evaluation and feedback, and explore the future of model fine-tuning.twitter.com/RazRazclehumanloop.comData Driven NYC YouTube Channeltwitter.com/mattturcklinktr.ee/mattturckShownotes: [00:00:47] How Humanloop helps product and engineering teams build reliable applications on top of large language models by providing tools to find, manage, and version prompts;[00:03:05] Where Humanloop fits into the MAD landscape as LM / LLM Ops;[00:02:40] The challenges of evaluating and monitoring LLM;[00:03:40] Why evaluating LLMs and generative AI is subjective given its stochastic attributes;[00:04:40] Why evaluation is important during development and production stages of LLMs to make informed design decisions, and how that challenge evolves In production to monitoring system behavior;[00:05:40] The need for regression testing with LLMs;[00:06:10] How Humanloop makes it easy for users to capture feedback including Implicit signals of user satisfaction, such as post-interaction actions and edits to generated content;[00:07:40] Why and how Humanloop uses guardrails in the app to ensure effective LLM use and implementation;[00:08:38] Why using an LLM as part of the evaluation process can introduce additional uncertainty and noise; with turtles all the way down;[00:09:40] How evaluators on Humanloop are restricted to binary yes-or-no style questions or numerical scores to maintain reliability with LLMs in production.[00:10:40] Why a new set of tools were needed to monitor and observe LLM performance;[00:11:40] How Humanloop’s interactive environment allows users to find and fix bugs in a prompt, including logs to support issue identification, and then run what-if style analysis by changing the prompt or information retrieval system — allowing for quick interventions and turnaround times within minutes to hours instead of days/weeks;[00:12:40] Why having evaluation and observability closely connected to prompt engineering tools is critical for speed;[00:13:40] How prompt engineering is like writing software specifications for the model, enabling domain experts to have a more direct impact on product development, and democratizing access and reducing reliance on engineers to implement the desired features;[00:15:40] The key differences between popular LLMs on the market today;[00:18:40] How the quality of open-source models has been rapidly improving, and how LLMs use tools or function calling to access APIs to go beyond simple text-based interactions;[00:21:22] How Humanloop empowers non-technical experts;[00:22:40] Where Humanloop fits within the AI ecosystem as an collaborative tool for enterprises building language models where collaboration and robust evaluation are crucial;[00:25:40] How Humanloop customers are often problem-aware, and how the go-to-market motion is mainly inbound, but sales-led[00:27:48] How Humanloop serves as a central place for storing prompts and sharing learnings across teams;[00:28:24] Raza’s thoughts on Open Source v. Closed Source models in the AI community;[00:30:40] The potential consequences of restricting access to models and Raza’s case for regulating end use cases and punishing malicious use rather than banning the technology altogether;[00:33:40] Next steps for Humanloop;

Today we're joined by Akilesh Bapu, CEO and Founder of DeepScribe, the platform using AI and Natural Language Processing to doctor/ patient transcripts. Matt and Akilesh go into DeepScribe's clinical use cases, supervised vs. unsupervised learning, and how critical it still is to have a human in the loop in a medical setting.

Today we have the pleasure of chatting with Sharon Zhou, CEO of Lamini, an LLM platform for the enterprise. Matt and Sharon go over the battle between prompting and fine-tuning, how the Lamini platform enables fine-tuning to be done "one billion times faster", and their recently-announced "LLM Super-station" in partnership with AMD.