Oct 3, 2024 · 53m · mad
Building The Database That Can Do It All | Tobie Morgan Hitchcock, CEO of SurrealDB
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of The MAD Podcast, host Matt Turck interviews Tobie Morgan Hitchcock, CEO and Co-Founder of SurrealDB, exploring the seven-year journey of building a unified, multi-model database engine in Rust to replace fragmented data backends, integrate AI, and streamline cloud infrastructure.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 22.2% of the talking time here. How this is scored →
Matt as informed peer 3.5 Guest teaching 5.2 Guest disagreement 0.9 Matt pushing back 1.8
speaking balance: gold is Matt, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 43:01 Rejecting Postgres-Based Architecture Limits
Tobie explicitly rejects the modern trend of building databases on top of Postgres, arguing that Postgres-based competitors are fundamentally constrained by Postgres limitations.
Hardest push from Matt ▶ 26:24 Pressing on Multi-Model Performance Trade-offsMatt refuses to accept multi-model marketing claims at face value, directly demanding to know where the performance catch is relative to specialized best-of-breed databases.
Biggest teaching moment ▶ 20:50 Explaining Record ID Architecture Over JoinsTobie educates the host on SurrealDB's core architectural breakthrough of record IDs, demonstrating how table-prefixed pointer arrays eliminate joins and expensive index scans.
Matt holds his own ▶ 19:30 Framing the 15-Year Database Unbundling ThesisMatt demonstrates deep enterprise software expertise by framing a 15-year history of database unbundling and challenging Tobie to prove why market rebundling is technically possible.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Matt as informed peer | Guest teaching | Guest disagreement | Matt pushing back | Why |
|---|---|---|---|---|---|---|
| Welcome and Tobie Hitchcock's Podcast Debut | 2 | 3 | 0 | 1 | Matt warmly welcomes Tobie on his podcast debut and asks for background on SurrealDB's early origins. Tobie clarifies that although people assume two founders wrote the codebase, he wrote all the code himself over seven years while his brother handled branding. | |
| Rethinking Database Architecture and Master's Thesis | 2 | 5 | 1 | 1 | Tobie explains his master's thesis background in key-value stores and temporal querying. He educates the host on treating time as a third dimension in database architecture to enable historical point-in-time state analysis. | |
| The Radical Codebase Rewrite into Rust | 3 | 5 | 1 | 1 | Matt references the industry adage that building a database takes ten years. Tobie details struggling against the Golang compiler and parser limitations before making the bold decision to rewrite the whole codebase in Rust. | |
| Current Milestones and Multi-Model Core Concepts | 3 | 6 | 1 | 1 | Matt highlights the distinction between multi-model and multi-modal definitions in AI contexts. Tobie walks through how SurrealDB unifies document, graph, and time-series paradigm access without traditional SQL joins. | |
| Separating Compute and Storage for Real-Time Analytics | 2 | 6 | 0 | 0 | Tobie breaks down how SurrealDB functions as a hybrid transactional analytical system by decoupling storage from compute. He demonstrates how this allows real-time analytical queries across graph and document records. | |
| Rebundling Database Markets and Record ID Innovations | 6 | 6 | 2 | 4 | Matt presents a structured market thesis arguing that database history is defined by 15 years of unbundling, challenging Tobie on why rebundling is possible. Tobie schools the host on record IDs, explaining how table-prefixed pointer arrays bypass traditional index lookups and joins. | |
| Performance Trade-offs and Benchmark Standards | 5 | 5 | 2 | 6 | Matt refuses to accept multi-model claims without scrutinizing trade-offs, directly asking Tobie where the catch is compared to best-of-breed single-purpose tools. Tobie acknowledges niche edge cases where specialized in-memory engines win while criticizing gamed industry benchmarks. | |
| AI Integration: Bringing Models to the Data | 4 | 5 | 0 | 1 | Matt asks about SurrealDB's role in the AI infrastructure stack. Tobie explains hosting machine learning models directly inside the database engine, allowing predictions on GPU hardware adjacent to data rather than piping records to microservice pipelines. | |
| Real-World Customer Adoption and Use Cases | 4 | 5 | 1 | 2 | Matt references pre-show context regarding e-commerce product recommendation search. Tobie clarifies enterprise adoption patterns, noting companies adopt SurrealDB to resolve targeted query bottlenecks rather than performing immediate full-stack database replacements. | |
| What's New in SurrealDB 2.0: Security, GraphQL, and Time Travel | 3 | 6 | 0 | 0 | Matt asks for major highlights in SurrealDB 2.0. Tobie details programmatic query loops, security authorization tokens, auto-generated GraphQL schemas, and temporal time-travel capabilities across connected graph models. | |
| Surreal Cloud Launch and Storage Architecture | 2 | 6 | 0 | 0 | Matt inquires about the Surreal Cloud availability launch. Tobie explains their cloud storage architecture, which decouples compute, caching nodes, and S3 object storage to deliver scalable auto-recovery and high durability. | |
| Navigating the Competitive Database Landscape | 5 | 5 | 3 | 4 | Matt presses Tobie on standing out in a crowded database market dominated by modern Postgres-based wrappers. Tobie forcefully reframes the market, arguing Postgres wrappers remain constrained by legacy Postgres limitations whereas SurrealDB built a fresh multi-model engine from scratch. | |
| The Founder Journey and Co-Founder Dynamics | 4 | 4 | 1 | 2 | Matt highlights startup founder statistics and historical parallels like Stripe's Collison brothers. Tobie discusses transition challenges, managing community expectations, and establishing absolute division of labor and trust when co-founding with his brother. |
Statements from this episode (20)
Hitchcock: SurrealDB reduces developer costs by consolidating database types
“So CeruleDB is reimagining databases, so effectively the tool that people use to store data, and then once they've stored it, to query that data. The thing about CeruleDB is it makes it much simpler in terms of querying that data, and it means that you can Red…”
Hitchcock: Firebase is a data platform, not a true database
“It's not really a database, it's what I would call a data platform or a data layer.”
Hitchcock: SurrealDB is built entirely in Rust
“CeruleDB, the entirety of CeruleDB is built in Rust.”
Hitchcock developed the SurrealDB codebase alone over seven years
“So it's just one person, and it was seven years in the making,”
Hitchcock: Traditional database engineers rarely rewrite backends from scratch
“If you were to have come from a database background, you probably wouldn't say, yup, we'll rewrite the entire database backend.”
Hitchcock: No existing database was designed for historical temporal querying
“There's not a single, really a single database out there designed in that specific way to be able to do that”
Hitchcock: I would not have started SurrealDB knowing it takes ten years
“Probably not. Definitely not. No, definitely not.”
Hitchcock: SurrealDB completely rewrote its original codebase from Golang to Rust
“And so, we made the decision to completely rewrite the entire codebase from Golang into Rust.”
Hitchcock: Rewriting a database entirely in Rust beats incremental migration
“What you'll see with other databases is they then design parts of it in Rust, and then gradually more and more that becomes Rust over time, and maybe some of it doesn't, but it's much easier when you go forward from the basis of everything is built in Rust.”
Hitchcock: SurrealDB runs embedded, single-node, or as a distributed cluster
“So SurrealDB can run as an embedded database, like SQLite. It can run as a single node, like Postgres. Or it can run as a distributed cluster, and you can scale the storage layer, and you can scale the compute layer.”
Hitchcock: Managing multiple specialized database backends degrades overall system performance
“And actually, if you're doing that, With three or four different database backends that can become incredibly complicated, The performance can degrade over time as well, or just be harder to reach because you're working with different platforms.”
Hitchcock: Dedicated in-memory graph databases will always outpace multi-model engines
“If you're dealing with graph data that all sits inside memory and it's being stored in a graph data structure, then that's always going to be quicker on a graph database that is sitting on a single node.”
Hitchcock: SurrealDB allows users to run AI models directly inside databases
“One of the pieces of functionality we have in CeruleeB is the ability to bring a model right inside the database so that you can run that model. It's a custom trade model or an off-the-shelf model. You can run that model right alongside your data.”
Hitchcock: Microservices simplify initial code but introduce massive systemic complexity
“Everyone built microservice-based architectures because that was the easiest thing to do. It meant that you could keep components of your code small, and you could, you build up your application from there. But it also led to a large, massive amount of complex…”
Hitchcock: Enterprises adopt SurrealDB for isolated problems, not to replace systems
“They're not coming to us and saying, I'm going to replace four different databases that I've spent the last 10 years implementing. It's usually being able to solve one problem to begin with, and that problem is usually that they're having to work with many dif…”
Hitchcock: SurrealDB will be the first database with multi-model time travel
“We will be the first database that enables you to time travel in your data in the different types of models that you can work with in SurrealDB.”
Hitchcock: SurrealDB avoids PostgreSQL limitations by not building on top of it
“We're not building on top of Postgres and therefore limited by what Postgres has to offer.”
Hitchcock: Most database startups focus on incremental fixes over reimagining architecture
“A lot of people, as I said at the very beginning, you know, a lot of organizations are focused on improving the performance of an analytical database, or reducing the cost of an analytical database, or adding some functionality to an alternative product that's…”
Hitchcock: Users expect new databases to immediately match twenty-year-old incumbents
“The outside world can love you and hate you at the same time. So they can love what you're building, but at the same time they expect you to be where a 20 year old database is in terms of functionality or performance, and I think sometimes people forget that t…”
Hitchcock: Technical startups cannot rely on engineering quality without brand identity
“In any organization, but also importantly in a startup as well, you can't just focus on the brand, or you can't just focus on the tech, you have to focus on both. For Surreal to be successful, we have to be an incredible technological product, and we have to h…”