I don't get crushed. I crush myself.

1/12. I'm excited to share our latest technical blog post on ParadeDB. After a brief hiatus focused on transforming ParadeDB into an enterprise-ready database, expect to hear a lot more from us. Today's post: How ParadeDB built an LSM on top of Postgres block storage. 🧵

LLMs don’t make mistakes. They do exactly what they were programmed/trained to do.

Let’s be clear: keeping amortization requirements for international R&D under Section 174 offers no tangible benefits to US businesses, bordering on being globally and domestically less competitive – short of closing the US market to non-US companies. So, whom does it benefit?

People never learn…

Next week, July 17th, I am hosting a database roundtable dinner in SF – for engineers who are interested in advancing the state of the art of core database stacks. No VCs. We have a few spots available – reach out if you're interested in joining!

We've forgotten Tcl/Tk unfairly. Also, 9.0 is out.

OLAP is always late, but can tell a good story.

One day, people will be like "wait, I think this is human slop – you wrote this without AI!.. Tsk tsk!" "I can spot human writing"

I did the unspeakable. I imagined that we went back in time, pre-LLM hype. And I tried to compare it to today to see if anything got better or worse because of it. The results were eye-opening.

When you prompt a model, it's largely ephemeral. But what it produces can and will prompt you, and you are not stateless.

I don’t want productivity. I want clarity and agency.

Last night was a blast – a fantastic group of database engineers and researchers got together to talk shop and have fun! Over three hours of flowing simultaneous conversations – it was really hard to wrap it up. The idea was simple – swap a Michelin-star private room dinner for

There is no AI slop. AI slop is the ultimate reflection of our own slop.

Building human-first companies, not as a virtue, but because clarity and agency make them unbeatable.

Imagine writing thoughtful blog posts, a book or a paper, only to be processed by a large language model to be summarized. Reading what the author put time, years and a part of their identity is the most respectful thing to do to keep them going.

LLM productivity is often antithetical to "less is more"

Woke up to a much more reduced mobility due to pain today. Being a runner, this is rather non-ideal.

Fun stuff – continuous SQS consumption into a Postgres table.

If you treat your database as permanent and external services as transient, you are already practicing Data-Native Architecture (DNA).