@nateberkopec.bsky.social
Prosopite + LLMs create a powerful workflow automation where you can get a PR fixing an N+1 with zero humans writing code
October 29, 2025 at 5:03 PM
Prosopite + LLMs create a powerful workflow automation where you can get a PR fixing an N+1 with zero humans writing code
this logo goes so hard
October 23, 2025 at 1:26 AM
this logo goes so hard
You won't say it, but I will
October 19, 2025 at 11:38 PM
You won't say it, but I will
Sorry you got filtered by memorizing one (1) bit of arbitrary information, I guess minitest is not for you
October 5, 2025 at 8:38 PM
Sorry you got filtered by memorizing one (1) bit of arbitrary information, I guess minitest is not for you
I made a utility to combine the llm CLI tool with Monologue's voice transcription by Every.
Brain-dump the WHY using your voice, let the LLM summarize the WHAT using git diff, and have it combine both into a complete, useful commit message.
Brain-dump the WHY using your voice, let the LLM summarize the WHAT using git diff, and have it combine both into a complete, useful commit message.
October 2, 2025 at 10:45 PM
I made a utility to combine the llm CLI tool with Monologue's voice transcription by Every.
Brain-dump the WHY using your voice, let the LLM summarize the WHAT using git diff, and have it combine both into a complete, useful commit message.
Brain-dump the WHY using your voice, let the LLM summarize the WHAT using git diff, and have it combine both into a complete, useful commit message.
I had three people from Brazil on a call recently because they wanted to start contributing to Puma but were unsure of where to start. This is something I've done for years and I've always loved it. We sometimes get really important new contributors from these convos.
October 1, 2025 at 5:00 PM
I had three people from Brazil on a call recently because they wanted to start contributing to Puma but were unsure of where to start. This is something I've done for years and I've always loved it. We sometimes get really important new contributors from these convos.
Some of the code these AI models put out is truly an abomination
October 1, 2025 at 6:01 AM
Some of the code these AI models put out is truly an abomination
That is not accurate re: commits
September 30, 2025 at 9:55 PM
That is not accurate re: commits
I absolutely love that the second point in Rust's CoC is "stop being so horny in project spaces". We love you, Rustaceans. Without you, OSS would be so much more boring.
September 29, 2025 at 11:30 PM
I absolutely love that the second point in Rust's CoC is "stop being so horny in project spaces". We love you, Rustaceans. Without you, OSS would be so much more boring.
If you have not put jemalloc on all your production Ruby systems yet, you should probably clear your schedule.
September 29, 2025 at 5:04 PM
If you have not put jemalloc on all your production Ruby systems yet, you should probably clear your schedule.
In what sense is/was rubygems and bundler in the "portfolio" of Sam G and Andre? This pitch reads like: "this critical project is unsustainable unless it is maintained by ME with YOUR MONEY".
September 29, 2025 at 4:08 AM
In what sense is/was rubygems and bundler in the "portfolio" of Sam G and Andre? This pitch reads like: "this critical project is unsustainable unless it is maintained by ME with YOUR MONEY".
more context from lobsters. so even the rv stuff at the end is ticky-tack bullshit that no one involved was actually mad about 🤷♂️
September 29, 2025 at 1:14 AM
more context from lobsters. so even the rv stuff at the end is ticky-tack bullshit that no one involved was actually mad about 🤷♂️
Normally Datadog is an absolute moneypit, but their Database Monitoring product is an incredible value at $70/mo.
It's a supercharged AWS Performance Insights dashboard, but far more useful and integrated with the rest of DD. Price doesn't change if you've got a 192 core DB.
It's a supercharged AWS Performance Insights dashboard, but far more useful and integrated with the rest of DD. Price doesn't change if you've got a 192 core DB.
September 24, 2025 at 5:04 PM
Normally Datadog is an absolute moneypit, but their Database Monitoring product is an incredible value at $70/mo.
It's a supercharged AWS Performance Insights dashboard, but far more useful and integrated with the rest of DD. Price doesn't change if you've got a 192 core DB.
It's a supercharged AWS Performance Insights dashboard, but far more useful and integrated with the rest of DD. Price doesn't change if you've got a 192 core DB.
I used to have some wolfman bags that looked very similar. I always liked the bumblebee coloring.
September 7, 2025 at 11:44 PM
I used to have some wolfman bags that looked very similar. I always liked the bumblebee coloring.
Despite all the improvements that have been landed on this stuff over the years, logging the source code line for every SQL query remains pretty expensive. It's better to use your APM for this, which will only add it to ~1% of your transactions, rather than 100%.
August 21, 2025 at 5:01 PM
Despite all the improvements that have been landed on this stuff over the years, logging the source code line for every SQL query remains pretty expensive. It's better to use your APM for this, which will only add it to ~1% of your transactions, rather than 100%.
I've been using this combination of prompt+tool to get Claude to setup arbitrary Rails apps dev envs for me. It's got a pretty good success rate, I would say about 80%.
August 8, 2025 at 5:02 PM
I've been using this combination of prompt+tool to get Claude to setup arbitrary Rails apps dev envs for me. It's got a pretty good success rate, I would say about 80%.
From the Speedshop skunkworks laboratory: effects of RUBY_THREAD_TIMESLICE.
In this example, we hammer a 10 thread Puma server with a high IO workload. Note how drastically the amount of time spent waiting for the GVL shifts around and how long the spans are.
In this example, we hammer a 10 thread Puma server with a high IO workload. Note how drastically the amount of time spent waiting for the GVL shifts around and how long the spans are.
August 4, 2025 at 4:57 PM
From the Speedshop skunkworks laboratory: effects of RUBY_THREAD_TIMESLICE.
In this example, we hammer a 10 thread Puma server with a high IO workload. Note how drastically the amount of time spent waiting for the GVL shifts around and how long the spans are.
In this example, we hammer a 10 thread Puma server with a high IO workload. Note how drastically the amount of time spent waiting for the GVL shifts around and how long the spans are.
LLMs have me thinking about this chart almost daily. We are in the age of self-ware. Small throwaway utility software for an audience of one.
August 1, 2025 at 4:56 PM
LLMs have me thinking about this chart almost daily. We are in the age of self-ware. Small throwaway utility software for an audience of one.
I wish everyone had this tool that my clients Pronto built: temporarily set up a production db copy.
They take a snapshot of prod every day. This page allows a dev to start a new db instance, same size, same data. Fire up psql and test all the indexes you want w/o fear.
They take a snapshot of prod every day. This page allows a dev to start a new db instance, same size, same data. Fire up psql and test all the indexes you want w/o fear.
July 31, 2025 at 5:04 PM
I wish everyone had this tool that my clients Pronto built: temporarily set up a production db copy.
They take a snapshot of prod every day. This page allows a dev to start a new db instance, same size, same data. Fire up psql and test all the indexes you want w/o fear.
They take a snapshot of prod every day. This page allows a dev to start a new db instance, same size, same data. Fire up psql and test all the indexes you want w/o fear.
I've always been a "gnomish" programmer and I think this technology is really designed for gnomes
en.wikipedia.org/wiki/Wikipe...
en.wikipedia.org/wiki/Wikipe...
July 30, 2025 at 4:56 PM
I've always been a "gnomish" programmer and I think this technology is really designed for gnomes
en.wikipedia.org/wiki/Wikipe...
en.wikipedia.org/wiki/Wikipe...
Claude wrote two commits in Go for me, a language that I do not understand the first thing about, and those commits were accepted in the OSS project Thruster:
July 28, 2025 at 4:56 PM
Claude wrote two commits in Go for me, a language that I do not understand the first thing about, and those commits were accepted in the OSS project Thruster:
A lot of people in our CGRP Slack community are having issues with Router 2.0 on Heroku. I initially thought this was more Puma/keepalive problems but actually we're seeing it with Falcon, Anycable, and Puma with keepalives disabled too!
July 22, 2025 at 9:11 PM
A lot of people in our CGRP Slack community are having issues with Router 2.0 on Heroku. I initially thought this was more Puma/keepalive problems but actually we're seeing it with Falcon, Anycable, and Puma with keepalives disabled too!
Native structured logging PR proposal for Rails, extracted from Shopify. I'm not a huge fan of the gems available for this currently, so first class support in the framework itself (fiber-aware, phew!) would be a huge plus
July 14, 2025 at 8:23 PM
Native structured logging PR proposal for Rails, extracted from Shopify. I'm not a huge fan of the gems available for this currently, so first class support in the framework itself (fiber-aware, phew!) would be a huge plus
I keep coming back to this, from Birgitta via Martin Fowler's recent post on LLMs - we are entering the age of non-deterministic development.
July 10, 2025 at 5:02 PM
I keep coming back to this, from Birgitta via Martin Fowler's recent post on LLMs - we are entering the age of non-deterministic development.