Command Prompt, Inc.

10/15/2026

AGLedger started with both Valkey and PostgreSQL. Once PostgreSQL could handle almost everything Valkey was doing, keeping a second database became difficult to justify.

In Part 2 of "When Postgres Is Enough", Michael Cooper explains how PostgreSQL gave him the confidence he needed to run his business.

Watch Part 2: https://youtu.be/_rLZQJIqMAs

In Part 2 of "When Postgres Is Enough", Michael Cooper and JD exami...

10/14/2026

When custom numeric logic doesn't fit cleanly into a vectorized NumPy operation, performance in standard Python loops degrades quickly.

Session 4 of Valerii’s Cython mini-series compares a Python loop, a Cython memoryview, and NumPy’s built-in clip to show where Cython actually fits.

Cython shouldn’t replace working NumPy code; it is the tool for custom loops that NumPy cannot vectorize.

Learn more on YouTube at

NumPy is already fast when a problem can be expressed with vectoriz...

10/14/2026

The value of AI isn't measured by how much work you hand over to it.

SysAdmin James Campbell shares why engineering judgment, creativity, and experience still matter.

Watch the full Group Swim discussion:
https://youtu.be/srFsNtoCLF4

10/13/2026

How much change do you want from a database you already trust? PostgreSQL’s cautious approach helps keep application behavior predictable, but it can also leave developers waiting for changes they need.

For James Nelson, Founder and CTO of PgCache, that reliability is what he loves about Postgres. In this T3D Postgres Love Letter, he and JD consider the friction that comes with it and how forks create room to experiment beyond what core PostgreSQL supports.

Watch the conversation: https://youtu.be/iBtHtSzsZBw

10/08/2026

If AI makes software easier to build, why would customers still pay for a software product?
Lev Kokotov is building a business around PgDog, an open source tool for scaling PostgreSQL. He uses AI every day, but the code is only part of what his customers need. Someone still has to keep the software working when their business depends on it.

JD and Lev discuss the question from database operations to the future of software businesses and work itself.

Watch the full conversation:

If AI makes software easier to build, what makes it worth paying fo...

10/06/2026

Even compiled Cython code degrades to standard Python speeds if the hot path remains tethered to the interpreter.

Session 3 of our Cython mini-series uses the annotation report to expose these interaction points instead of guessing. We benchmark each change with timeit and migrate data to a typed memoryview to resolve the bottleneck.

A compiled Cython module is not necessarily an optimized one.In S...

10/06/2026

How current do your PostgreSQL materialized views need to be?
For a live scoring site, waiting for the next refresh means the displayed score can fall behind as people submit new results.

Brian Fehrle, DBA at Command Prompt, wants real-time materialized views. He’s built trigger-based versions for simpler cases, but aggregates make that work more involved.

In this Postgres Wishlist episode, he and JD consider what implementation might require, along with a related request: refreshing just the relevant results when customers have new activity.

Watch the T3D conversation:

What would it take for PostgreSQL materialized views to stay curren...

10/05/2026

Do you ever feel like you're trying to solve stress with something complicated?

Amanda Nystrom and JD wrap up their stress management series with a simple reminder: sleep, movement, boundaries, connection, and other everyday habits often have a bigger impact than we think.

Listen to the full conversation:

In this episode of Health and Tech, Amanda Nystrom and JD wrap up t...

10/01/2026

Compiling a .pyx file is not the same as optimizing the code inside it, and does not resolve Python overhead if the code retains Python-level object semantics.
Session 2 walks through building and importing a Cython modulesetup.py and pyximport. We demonstrate why strict type declarations are required to bridge the gap between a compile and a performance gain.

https://youtu.be/t18_P6F-vqU

09/29/2026

Could PostgreSQL’s logical replication stream tell us more about schema changes? That’s James Nelson, Founder and CTO of PgCache’s Postgres Wish: a lightweight way to receive information about DDL.

In this T3D episode, James joins JD to consider what that would involve. The conversation covers the appeal of logical replication’s selectivity, the overhead of a cluster-wide stream, and whether triggers and custom messages could offer a workaround.

Watch the full conversation: https://youtu.be/YC5B1zjFaZ0

PostgreSQL logical replication can provide a selective stream of da...