The MAD Podcast with Matt Turck

Before he founded Modal, Erik Bernhardsson created Spotify's music recommendation system. Today he's bringing a consumer app approach to radically simplifying developer experience for data and AI projects on the Modal platform.

In this episode, we dive into the broader AI compute landscape, discussing the roles of hyperscalers, GPU clouds, inference platforms, and the emergence of alternative AI cloud providers. Erik gives us a product tour of the Modal platform, provides insights into the AI industry's shift from training to inference as the primary use case, and speculates on the future of AI-native consumer applications. Learn about Modal's commitment to fast feedback loops, their cloud maximalist approach, their dedication to building a product that developers truly love, as well as founder lessons Erik learned along the way.

Erik's blog: https://erikbern.com

"It's hard to write code for humans": https://erikbern.com/2024/09/27/its-hard-to-write-code-for-humans

Modal

Website - https://modal.com

Twitter - https://x.com/modal_labs

Erik Bernhardsson

LinkedIn - https://www.linkedin.com/in/erikbern

Twitter - https://x.com/bernhardsson

FIRSTMARK

Website - https://firstmark.com

Twitter - https://twitter.com/FirstMarkCap

Matt Turck (Managing Director)

LinkedIn - https://www.linkedin.com/in/turck/

Twitter - https://twitter.com/mattturck

(00:00) Intro

(01:35) What is Modal?

(02:18) Current state of AI compute space

(09:54) Erik's path to starting Modal

(13:57) Core elements of the Modal platform

(28:52) Is serverless the right level of abstraction for AI compute?

(33:35) Balancing costs: GPU vendor fees vs. customer pricing

(37:56) Designing products for humans

(42:43) Modal's early go-to-market motion

(45:32) Managing early engineering team

(48:26) The only correct way to add a new function to the company

(50:07) Building company in NYC

(52:05) Modal's roadmap

(54:04) Erik's predictions on AI

A founding engineer on Google BigQuery and now at the helm of MotherDuck, Jordan Tigani challenges the decade-long dominance of Big Data and introduces a compelling alternative that could change how companies handle data.

Jordan discusses why Big Data technologies are an overkill for most companies, how MotherDuck and DuckDB offer fast analytical queries, and lessons learned as a technical founder building his first startup.

Watch the episode with Tomasz Tunguz: https://youtu.be/gU6dGmZzmvI

Website - https://motherduck.com

Twitter - https://x.com/motherduck

Jordan Tigani

LinkedIn - https://www.linkedin.com/in/jordantigani

Twitter - https://x.com/jrdntgn

FIRSTMARK

Website - https://firstmark.com

Twitter - https://twitter.com/FirstMarkCap

Matt Turck (Managing Director)

LinkedIn - https://www.linkedin.com/in/turck/

Twitter - https://twitter.com/mattturck

(00:00) Intro

(00:56) What is the Small Data?

(06:56) Marketing strategy of MotherDuck

(08:39) Processing Small Data with Big Data stack

(15:30) DuckDB

(17:21) Creation of DuckDB

(18:48) Founding story of MotherDuck

(24:08) MotherDuck's community

(25:25) MotherDuck of today ($100M raised)

(33:15) Why MotherDuck and DuckDB are so fast?

(39:08) The limitations and the future of MotherDuck's platform

(39:49) Small Models

(42:37) Small Data and the Modern Data Stack

(46:47) Making things simpler with a shift from Big Data to Small Data

(50:04) Jordan Tigani's entrepreneurial journey

(58:31) Outro

With a $4.5B valuation, 5M AI builders and 1M public AI models, Hugging Face has emerged as the key collaboration platform for AI, and the heart of the global open source AI community.

In this episode of The MAD Podcast, we sit down with Clément Delangue, its co-founder and CEO, and delve deep into Hugging Face's journey from a fun chatbot to a central hub for AI innovation, the impact of open-source AI and the importance of community-driven development, and discuss the shift from text to other AI modalities like audio, video, chemistry, and biology. We also cover the evolution of Hugging Face's business model, and the different approach to company culture that the founders have implemented over the years.

Hugging Face

Website - https://huggingface.co

Twitter - https://x.com/huggingface

Clem Delangue

LinkedIn - https://www.linkedin.com/in/clementdelangue

Twitter - https://x.com/clemdelangue

FIRSTMARK

Website - https://firstmark.com

Twitter - https://twitter.com/FirstMarkCap

Matt Turck (Managing Director)

LinkedIn - https://www.linkedin.com/in/turck/

Twitter - https://twitter.com/mattturck

(00:00) Intro

(01:46) Miami vs. New York vs. San Francisco

(03:25) Current state of open source AI

(11:12) Government regulation of AI

(13:18) What is open source AI?

(15:21) Open source AI: China vs U.S.

(18:32) LLMs vs. SLMs

(22:01) Are commercial LLMs just 'Training Wheels' for enterprises?

(24:26) Software 2.0: built with AI

(28:03) Hugging Face founding story

(37:03) Are there any competitors?

(44:06) Most interesting models on Hugging Face

(50:35) Shifting focus in enterprise solutions

(55:06) Bloom & Idefix

(58:44) The culture of Hugging Face

(01:04:44) The future of Hugging Face

This episode is a captivating conversation with Richard Socher, serial entrepreneur, investor, and AI researcher.

Richard elaborates on why he likens the impact of AI to the Industrial Revolution, the Enlightenment, and the Renaissance, discusses important current issues in AI, such as scaling laws and agents, provides a behind-the-scenes tour of YOU.com and its evolving business model, and finally describes his current investment strategy in AI startups.

You.com

Website - https://you.com/business

Twitter - https://x.com/youdotcom

Richard Socher

LinkedIn - https://www.linkedin.com/in/richardsocher

Twitter - https://x.com/richardsocher

FIRSTMARK

Website - https://firstmark.com

Twitter - https://twitter.com/FirstMarkCap

Matt Turck (Managing Director)

LinkedIn - https://www.linkedin.com/in/turck/

Twitter - https://twitter.com/mattturck

(00:00) Intro

(02:00) "AI era is the Industrial Revolution, Renaissance, and the Enlightenment combined"

(07:49) Top-performers in the Age of AI

(11:15) Comeback of the Renaissance Person

(13:05) People tried to stop Richard from doing deep learning research. Why?

(14:34) Jevons paradox of intelligence

(17:08) Scaling Laws in Deep Learning

(23:23) Can Deep Learning and Rule-Based AI coexist?

(25:42) Post-transformers AI Architecture

(28:20) Achieving AGI and ASI

(36:43) AI for everyday tasks: how far is it?

(44:50) AI Agents

(55:45) Evolution of You.com

(01:02:11) Technical side of You.com

(01:06:46) Is AI getting cheaper?

(01:13:05) What is AIX Ventures?

(01:16:36) VC landscape of 2024

(01:24:31) Research vs Entrepreneurship

(01:26:12) OpenAI’s transformation and its impact on the industry

In this episode, we sit down with Tobie Morgan Hitchcock, the founder of SurrealDB, to dive deep into the evolving world of databases and the future of data storage, querying, and real-time analytics.

(00:00) Intro
(02:03) What is SurrealDB?
(02:53) How did SurrealDB get started?
(09:10) The Challenges of Building a Database from Scratch
(10:36) Why SurrealDB Chose Rust
(12:54) A Deep Dive into SurrealDB’s Unique Features
(19:30) Why Now?
(26:32) What Sets SurrealDB Apart from Other Databases
(30:01) SurrealDB’s Role in the Future of AI and Machine Learning
(32:45) Why Developers Are Choosing SurrealDB
(36:14) What’s New in SurrealDB 2.0?
(40:10) SurrealDB Cloud: Scalability Meets Simplicity
(42:21) How SurrealDB Fits into the Competitive Database Landscape
(45:37) Early Lessons from Building SurrealDB
(48:34) Co-Founding SurrealDB with His Brother

In this episode, we dive deep into the story of how Datadog evolved from a single product to a multi-billion dollar observability platform with its co-founder, Olivier Pomel. Olivier shares exclusive insights on Datadog's unique approach to product development—why they avoid the "Apple approach" of building in secret and instead work closely with customers from day one.

You’ll hear about the early days when Paul Graham of Y Combinator turned down Datadog, questioning their lack of a first product. Olivier also reveals the strategies behind their iterative product launches and why they insist on charging early to ensure they’re delivering real value.

The second half of the conversation is focused on all things AI and data at Datadog - the company's initial reluctance to use AI in its products, how Generative AI changed everything, and Datadog's current AI efforts including Watchdog, Bits AI and Toto, their new time series foundational model.

We close the episode by asking Olivier about his thoughts on the topic du jour: founder mode!

In this episode, we sit down with Ali Dasdan, CTO of ZoomInfo, a titan in the B2B sector, who harnesses vast datasets and advanced AI to redefine sales and marketing for over 35,000 global customers with $21.2 billion in annualized revenue.

We delve deep into ZoomInfo's AI initiatives, including their transformative 'Copilot,' explore sophisticated data management, and discuss their dual platforms catering to internal and customer-facing needs.

ZoomInfo

Website - https://www.zoominfo.com

Twitter - https://x.com/zoominfo

Ali Dasdan

LinkedIn - https://www.linkedin.com/in/dasdan

Twitter - https://x.com/alidasdan

FIRSTMARK

Website - https://firstmark.com

Twitter - https://twitter.com/FirstMarkCap

Matt Turck (Managing Director)

LinkedIn - https://www.linkedin.com/in/turck/

Twitter - https://twitter.com/mattturck

(00:00) Intro

(02:03) What is ZoomInfo

(04:47) Data as service

(06:15) Ali Dasdan's story

(07:31) Organization of ZoomInfo

(10:48) ZoomInfo Data Platform

(21:02) Lessons from building a data platform

(23:19) AI application at ZoomInfo

(27:58) ZoomInfo's Copilot

(37:43) ZoomInfo AI toolstack

(39:30) Working with small vs. big companies in the AI business

(43:39) Using data and AI for internal productivity

In this episode, we sit down with Eric Glyman, co-founder of Ramp, the company that revolutionized finance management to become a powerhouse valued at $7.6 billion.

Eric shares the tradition of counting the days since Ramp's founding and how it fosters a sense of urgency and productivity, explains the use of AI to automate expense management and fraud detection, and gives an inside look at Ramp's cutting-edge AI products, including the Ramp Intelligence Suite and experimental agentic AI use cases.

Ramp

Website - https://www.ramp.com

Twitter - https://x.com/tryramp

Eric Glyman

LinkedIn - https://www.linkedin.com/in/eglyman

Twitter - https://x.com/eglyman

FIRSTMARK

Website - https://firstmark.com

Twitter - https://twitter.com/FirstMarkCap

Matt Turck (Managing Director)

LinkedIn - https://www.linkedin.com/in/turck/

Twitter - https://twitter.com/mattturck

(00:00) Intro

(01:49) What is Ramp?

(04:25) How did the company start?

(09:18) Technical aspects of Ramp infrastructure

(12:17) "We can tell you if you're paying too much"

(14:20) Data privacy at Ramp

(16:13) Data infrastructure tools used at Ramp

(17:58) Traditional AI use cases

(24:51) GenAI use cases

(27:47) AI/human interaction

(33:32) Ramp Intelligence Suite

(39:38) How Ramp keeps high product release and product velocity

(42:37) How did Ramp get to product-market fit?

(45:54) Eric's perspective on building a company in NYC

In this episode, we reconnect with Sharon Zhou, co-founder and CEO of Lamini, to dive deep into the ever-evolving world of enterprise AI.

We discuss how the AI hype is evolving and what enterprises are doing to stay ahead, break down the different players in the Inference market, explore how Memory Tuning is reducing hallucinations in AI models, the role of agents in enterprise AI, and the challenges of making them real-time and reliable.

Lamini

Website - https://www.lamini.ai

Twitter - https://x.com/laminiai

Sharon Zhou

LinkedIn - https://www.linkedin.com/in/zhousharon

Twitter - https://x.com/realsharonzhou

FIRSTMARK

Website - https://firstmark.com

Twitter - https://twitter.com/FirstMarkCap

Matt Turck (Managing Director)

LinkedIn - https://www.linkedin.com/in/turck/

Twitter - https://twitter.com/mattturck

(00:00) Intro

(02:18) The state of the AI market in July, 2024

(10:51) What is Lamini?

(11:43) What is Inference?

(15:36) GPU shortage in the enterprise

(18:06) AMD vs Nvidia

(22:10) What is Lamini's final product?

(25:30) What is Memory Tuning?

(29:01) What is LoRA?

(32:39) More on Memory Tuning

(35:51) Sharon's perspective on AI agents

(40:01) What is next for Lamini?

(41:54) Reasoning vs pure compute in AI

In this episode, we sit down with Jeremy Kahn, the AI Editor at Fortune Magazine, who has recently published a book called "Mastering AI: A Survival Guide to Our Superpowered Future".

Jeremy shares his unique insights on AI's potential risks and transformative benefits, including the importance of UI design in maximizing AI's utility, the potential for AI to create a "winner takes most" economy, and the need for thoughtful AI regulation to mitigate risks without stifling innovation.

Book: https://www.amazon.com/Mastering-AI-Survival-Superpowered-Future/dp/1668053322

Jeremy Kahn

LinkedIn - https://www.linkedin.com/in/jeremy-kahn-01100462

Twitter - https://x.com/jeremyakahn

FIRSTMARK

Website - https://firstmark.com

Twitter - https://twitter.com/FirstMarkCap

Matt Turck (Managing Director)

LinkedIn - https://www.linkedin.com/in/turck/

Twitter - https://twitter.com/mattturck

(00:00) Intro

(01:43) Why the UI design is important for AI?

(04:32) The book is called "Mastering AI". Why?

(12:03) Automation Bias vs Automation Surprise

(20:16) The role of AI in the future of science and art

(25:32) "I think mass unemployment is a red herring, but we might see a lot of disruption"

(34:19) Jeremy's perspective on Agentic AI

(36:29) Does AI development need to be regulated?

(38:56) Should we worry about the AGI and Superintelligence?

(42:18) Who provided the most thoughtful conversation for the book?

(43:57) "I didn't use AI for the book at all"

(46:20) Jeremy's work at Fortune