TL;DR
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
A Pragmatic Engineer episode features Peter Mattis discussing distributed databases, his work on Gmail and Google storage, and how AI has changed his coding work. The interview revisits engineering tradeoffs including B-trees and erasure coding, while Mattis argues that AI can help experienced engineers produce more code without lowering quality.
The Pragmatic Engineer has published an interview with Peter Mattis, co-founder and chief technology officer of Cockroach Labs, about distributed databases, storage systems and the effect of AI on software development. Mattis, who also helped create GIMP and worked at Google, describes engineering choices behind systems such as Gmail and Google’s file storage, and says AI has brought him back to writing code after his work shifted toward management.
The interview traces Mattis’s path from open-source software to Google and then Cockroach Labs. It recounts that Mattis and his college roommate Spencer Kimball created GIMP, and that an early version of the Google logo was made with the image editor. Mattis says he initially declined an offer from Sergey Brin in 2001 because he did not want to commute from San Francisco to Mountain View; he joined Google after the company contacted him again.
At Google, B-trees featured in several engineering efforts described in the episode. In Gmail’s early storage layer, the structures tracked email threads and unread counts. Mattis also built a B-tree intended to offer similar behavior to C++’s std::map, whose red-black-tree implementation appeared in memory profiles. The interview says the alternative used fewer pointers and improved spatial locality. Mattis later built a Swiss Table implementation for Go; the report says it was faster than the existing map implementation and that the Go team helped complete work that later entered the Go library.
The episode also covers Colossus, Google’s successor to the Google File System. The report says the system used Reed–Solomon erasure coding to store data in two copies rather than three full copies, reducing storage overhead by 33% while increasing redundancy. That figure and description are presented by the source as part of its account of Mattis’s work; the episode summary does not provide an independent technical comparison or further details about the calculation.
Storage Choices Behind Large Systems
The discussion connects data-structure choices to the practical demands of services that must store and retrieve information at scale. Gmail’s thread tracking and Google’s distributed file storage illustrate different layers of that work: one concerns organizing records for an application, while the other concerns storing files across a large system. Mattis’s examples show how engineers may revisit familiar structures when memory use or performance becomes a constraint.
The episode’s AI discussion is relevant to software teams weighing how coding work may change. Mattis says he feels more productive writing code with AI and sees potential for domain experts to increase their impact without a loss in quality. That is his experience and view, as presented in an interview; the source does not provide comparative productivity data or a study establishing that the same results apply across engineering teams.
Top picks for "distribut databas peter"
As an affiliate, we earn on qualifying purchases.
From GIMP to Cockroach Labs
Mattis’s career spans open-source software, Google and the company he co-founded, Cockroach Labs. The interview uses those experiences to frame a discussion of reliability and correctness in systems that operate at scale. It also returns to B-trees as a recurring tool: the source links them to Gmail’s early storage, Mattis’s alternative to std::map, and CockroachDB’s range index.
For Google file storage, the reported shift from three full copies in GFS to erasure coding in Colossus is an example of a design tradeoff: store less duplicated data while maintaining redundancy. Separately, the interview discusses network latency. It gives an example of round trips within a zone falling from milliseconds when Colossus was built to about 100 microseconds today. The source does not name a measurement method or specify the systems and conditions behind that comparison.
““There’s always going to be someone else working on your idea.””
— Peter Mattis, as quoted in The Pragmatic Engineer interview
Limits of the Interview’s Evidence
The material provided is an episode summary, not a full transcript, and it does not give the publication date. It also leaves open how Mattis measured his current productivity with AI, what tasks or time period he compared, and what evidence supports the claim of no decline in quality. The source offers no broader data on AI’s effects across software teams.
Some technical figures are presented without supporting methods in the supplied material. The 33% storage-overhead reduction and the latency comparison are attributed to the episode’s account, but readers are not given underlying measurements, operating conditions or links to technical documentation for those claims here. The summary also does not describe the specific architecture or tradeoffs behind CockroachDB’s current distributed database design.
Where to Hear the Discussion
The episode is available through YouTube, Apple and Spotify, according to The Pragmatic Engineer. The source says a transcript appears at the top of the episode page and timestamps are listed at the bottom. Readers seeking detail beyond the summary can consult those materials for Mattis’s fuller account of his engineering work and views on AI.
No product launch, company decision or follow-up announcement is identified in the supplied report. The next development for readers is access to the full interview and its transcript; any wider conclusions about AI productivity or the systems discussed would require evidence beyond the episode summary.
Key Questions
Who is Peter Mattis?
Peter Mattis is a co-founder and CTO of Cockroach Labs. He also helped create GIMP and worked at Google, including on Gmail and distributed storage, according to The Pragmatic Engineer.
What does the interview cover?
The episode discusses distributed databases, storage engineering, data structures and AI-assisted coding, drawing on Mattis’s work at Google and Cockroach Labs.
What did Colossus change about Google file storage?
The source says Colossus replaced three full copies of data with storage using Reed–Solomon erasure coding. It reports a 33% reduction in storage overhead while increasing redundancy, but the supplied summary does not include the underlying calculation.
What does Mattis say about AI and coding?
Mattis says AI has brought him back to writing code after his work shifted toward management. He believes it can raise productivity and help domain experts increase their impact without reducing quality; the source presents this as his view and experience, not as a measured industry-wide result.
Source: rss
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
