New post: "Make Worse Software, Slower" This is not satire. I argue for: – Rejecting functional programming – Using ORMs to perfectly hide impedance mismatches – Skipping event sourcing – Other proven best practices blog.redplanetlabs.com/2025/06/17/m...

I still see this argument sometimes in the Clojure community: 1. Specter is great at manipulating deeply nested data 2. You shouldn't nest data very deep to begin with 3. Therefore, Specter is useless Do you see the multiple flaws in the logic?

This talk begins with homoiconicity as a simple idea with huge impact ("opens up a new dimension of programming") that most programmers initially perceive as useless or harmful. It then explores four more simple ideas with that same dynamic that greatly simplify software dev.

Wow, there are a lot of gems in Alan Perlis's "Epigrams on Programming" from 1982

Here's the video of my keynote "Simple ideas with huge impact from Clojure and Rama" from reClojure two weeks ago. The ideas only take a few seconds to express, but they will seriously challenge how you perceive programming. www.youtube.com/watch?v=25lJ...

Arrived in London and looking forward to hanging out with the Clojure community here at reClojure!

Had a fun time showing my embarrassing editor setup and talking about Rama on the Apropos podcast www.youtube.com/watch?v=h8cd...

This is such a great sci-fi story

Published "Next-level backends with Rama: personalized content moderation in 60 LOC" This is a detailed line by line tutorial on using Rama for both Clojure and Java to implement a high-performance mute/unmute feature on a social network. blog.redplanetlabs.com/2025/04/29/n...

What I love most about this case study is how Rama enabled G+D Netcetera to greatly expand the market of their product. This is similar to what we saw in the Multiply case study: the mindset changes from "what does our tooling allow us to do?" to "what's our dream product?".

Very excited to publish this case study! What Rama did for this product is incredible: "How G+D Netcetera used Rama to 100x the performance of a product used by millions of people" blog.redplanetlabs.com/2025/04/22/h...

Since I'm so used to Clojure, what drives me nuts about writing Java code is the crazy number of parens you have to type

If you post anything about dynamic typing being superior to static typing, you'll soon find out from the incensed comments that many people don't know the difference between dynamic typing and weak typing and will fight you to the death over it.

The Rama case study we're publishing next week is a doozy. This customer operates a product used by millions of people, and Rama: - Improved the performance of their application on two critical metrics by over 100x - Reduced their AWS hosting costs by 55%

New post: "Next-level backends with Rama: fault-tolerant timed notifications in 25 LOC". This is a detailed tutorial showing how to manage future work in an application. Examples on this site are features like scheduled posts or poll expirations. blog.redplanetlabs.com/2025/04/16/n...

1/ Rama is uniquely able to build an entire backend at scale with diverse computation and storage needs. Some have misunderstood this to mean using Rama means you need to replace all existing infrastructure. That would be crazy.

Published "Next-level backends with Rama: recommendation engine in 80 LOC" This is a detailed line by line tutorial on using Rama for both Clojure and Java. As with all Rama apps, the example in the post is scalable, fault-tolerant, and ACID compliant. blog.redplanetlabs.com/2025/04/08/n...

This post is a demonstration of why colocation is superior to locking/transactions for atomicity

New post: "Massively scalable collaborative text editor backend with Rama in 120 LOC" blog.redplanetlabs.com/2025/04/01/m...

The way this does parallel graph traversals by recurring a loop multiple times per iteration while also jumping around the cluster is really cool. It's only a couple lines of code but exemplifies what I mean when I talk about Rama making "where" and "what" composable.

Published "Next-level backends with Rama: storing and traversing graphs in 60 LOC". This is a detailed line by line tutorial on using Rama for both Clojure and Java. As with all Rama apps, the example in the post is scalable, fault-tolerant, and ACID compliant. blog.redplanetlabs.com/2025/03/26/n...

Looking forward to speaking at reClojure in London on May 26th!

This is a reminder that writing state in global variables is a horrible thing to do, and databases are just global variables by a different name.

Now that the free version of Rama is out, you can directly test our claims of how much it improves backend development! I've found with our early users it takes a week or two to get through the initial learning curve.

Starting a new blog series next week called "Next-level backends with Rama" where each post is a detailed tutorial of using Rama for a specific use case, with all code in both Java and Clojure. The first post will be about storing and traversing graphs.

Also launched a major revamp for the Red Planet Labs website yesterday, with a lot more information on Rama and resources for learning redplanetlabs.com

Rama, the 100x developer platform, is now free for production use! You can download Rama from our website. The free version can run clusters up to two nodes, which is enough to comfortably run small/medium scale applications. More details in the blog post. blog.redplanetlabs.com/2025/03/18/r...

One day until we release the free version of Rama, along with a fresh new website and enhanced docs!

1/ Being a source of truth and being an indexed store that answers queries are fundamentally in conflict. It's impossible for one datastore to do both of these well.

Sign up to our mailing list to be notified as soon as we release the free version of Rama on March 18th redplanetlabs.us13.list-manage.com/subscribe?u=...

Looking forward to releasing the free version of Rama one week from today!

Published another Rama user case study: "How AfterHour built an ultra-scalable chat service in one month with Rama" blog.redplanetlabs.com/2025/03/11/h...

Good tips on learning Rama from an experienced Rama user

Published the first Rama user case study! "How Multiply went from Datomic to XTDB to Rama" blog.redplanetlabs.com/2025/03/04/h...

Rama will be free to use in three weeks on March 18th! Leading up to that, we'll be publishing our first two case studies detailing how Rama has dramatically improved the capability and development speed of our users.

I'll be keynoting reClojure in London on May 26th reclojure.org

In this talk I show how Rama's huge productivity gains come from enabling composability for backend development for both computation and storage www.youtube.com/watch?v=X__j...

1/ Released awesome backups feature for Rama today: - Takes a few seconds to configure - Backs up all state to an external store (e.g. S3) - Each backup run is incremental, only transferring new state since the last backup

I accidentally divulged the secrets of how we used Clojure to build Rama in this podcast. Please don't listen. www.youtube.com/watch?v=Exkz...

Looking forward to talking about Rama and its functional programming roots on Friday

This is Claude's reaction to my most recent blog post. Couldn't have said it better myself.

We released a new feature to Rama today I'm really excited about called "yielding select". The "API" for it is just a single option to an existing method, but underneath that option is an incredible amount of power that took months to implement. Summary: 🧵

The older I get, the more amazed I am that John McCarthy invented Lisp in 1958 when the popular programming languages were Algol, Fortran, and Cobol. 66 years later and it's still mind-blowing to the vast majority of the industry.

A traditional architecture of reading/writing to one or more databases from an app server is just stream processing with really bad fault-tolerance and tons of race conditions. If writing to multiple databases, it's nearly certain they will become inconsistent with each other.

My briefest summary of what makes Rama powerful: it makes all the computation and storage for a backend composable, which eliminates impedance mismatches and creates massive leverage.

Wielding randomness effectively in tests is by far the best defense against race conditions

I've encountered this only a handful of times, but there should be a name for a bug which, when fixed, reveals an even worse bug it was coincidentally suppressing. I propose "Pandora's bug".

It feels like half the work of supporting users of your software is helping with unrelated things, like OS configuration, using AWS/Azure, making build scripts, and so on.

I've noticed many functional and performance issues with @bsky.app. If only there was a fully-featured open-source implementation in a tiny amount of code that solves the challenges of running a Twitter-like service at scale with an unbalanced social graph. github.com/redplanetlab...

If you focus too much on preventing weak programmers from writing bad code, you're also going to prevent strong programmers from writing great code.