Josh Nichols

@nateberkopec Misread as “it will slop”, like the antonym of “it will slap”

this is exactly like the LLMs shipping that terrible SQLite implementation btw. when you put a token generator, human or LLM, and don't explicitly make it focus on a requirement (perf in this case), it will ship absolute slop

frontend developers will come to the office and talk your ear off about "craft" and "art" and then ship a 15mb react monstrosity that takes 13 seconds to LCP

~/.claude/ is basically production. No staging, no isolation. Every Claude Code plugin mutates your daily driver config. Wrote up how to set up a proper test environment after borking mine one too many times. bit.ly/4r7VlvC

@hakunin Oh no! I don’t think there will be anything I directly help, so definitely contact support if you haven’t already. Once you have a case number, you can send it my way and I’ll make sure it gets routed to the right place internally

@techpickles Hey Josh, fellow long time rubyist. I work at a company that's expecting W2Cs from Gusto after recent issues. It's really hard to get any insights into the state of things, whether there's any light at the end of the tunnel. Any chance we can dm?

@headius That'd be great! I'll DM to find a time to chat

@techpickles I'd love to have a chat about how JRuby could fit into all of that. The best of both worlds for scaling is to be able to handle requests and back-end calls asynchronously, while also parallelizing synchronous work. Let me know if we can set up a call or chat!

My team is hiring! Come work on sidekiq, karafka, and other async stuff with Gusto (see what I did there?) job-boards.greenhouse.io/gusto/jobs/741…

@bcardarella I’ve had that happen before when I said to use a subagent. I didn’t track it as a Task though, and just immediately came back said it is running in the background (?!)

I told Claude to one-shot an integration test against a detailed spec I provided. It went silet for about 30 minutes. I asked how it was going twice and it reassured me it was doing work. Then I asked why it was taking so long:

@nateberkopec The other tty- gems are great too. I’ve probably used tty-table and tty-markdown the most with it

Wow why have I never used tty-prompt? This is great. github.com/piotrmurach/tt…

@nateberkopec The same could be said of human-augmented dev, LLM just tends to have lot less context to start with

In the age of LLM-augmented dev, your job is to: 1. Put the AI in the right latent space (prompt) 2. Provide correct context 3. Refine the output with automated loops and deterministic tooling. 4. Review output and improve step 1-3 so that you dont provide the same review 2x.

@nateberkopec @bsilva96 The big one for me has been builtin support for things like aqua, ubi, etc that support installing stuff from GitHub releases. It’s useful even for packages you might normally install with homebrew, as you can tightly control the version.

@bsilva96 For me, I liked: parallel installs, less plugin management, easier to remember the commands, no shimming

All the smart people are saying it

@nateberkopec Ah yes, a fellow figlet enjoyer 🧐

Here's another one. Contract creation powered by pandoc:

agentic coding has really lowered my activation energy for writing new tools. I'm writing lots of small helpers and automating things that in the past would've taken way too long.

@nateberkopec I’ve ben using Cursor a lot lately, and it does a pretty good job getting context in a code base. I wonder how well it interpret profiles 🤔

Automatic profile reading doesn’t work because reading profiles requires asking: “what work here is necessary, and what is not?” That question requires context and experience with the codebase.

Stumbled across notes from my first @RenderATL (always keep your notes!). From @kishau : "Don't create glass ceilings for yourself by being an ineffective communicator and not being a good storyteller... If you make a technology choice, be able to talk through the alternatives you didn't pick and why" still true and relevant for my current role I didn't even have then🙏🏾

"None of our top 50 endpoints by throughput can have more than 5% of responses be >500ms" This is a good perf goal. It scopes to the most important parts of the app, is set based on expectations of human users (users can feel >500ms resp), and it's usually quite achievable.

I just released Zeitwerk 2.7. Now, you can define namespaces with manual constant assignments too (think, Namespace = Struct.new). To the best of my knowledge, this completes the most important goal of the gem: to fully match Ruby. 🎉 Let me elaborate a bit on why that was not supported before, and how we have been able to remove that long-standing limitation. First, let's take a step back to understand a tricky aspect of namespaces. Consider this Hotel namespace (in Ruby, both classes and modules can be namespaces): # hotel.rb class Hotel include Pricing end # hotel/pricing.rb module Hotel::Pricing # Pricing related logic extracted to a mixin. end You cannot load any of the two files in regular Ruby, right? To load hotel.rb you need Hotel::Pricing in place, but if you try to load hotel/pricing.rb, you need Hotel! But that works if the project is managed by Zeitwerk, and it worked with the classic autoloader too. To support that use case, the :class event of TracePoint was instrumental, and it was the only technique I could think of in 2018. It worked this way: When TracePoint notifies that Hotel has been created, the loader scans its subdirectories and defines autoloads (issues Module#autoload calls) for child constants, in particular for Hotel::Pricing. The key observation is that happens precisely before the include line is executed. When the interpreter hits that line, the autoload is in place and autoloading hotel/pricing.rb is possible (because Hotel exists!). Problem was, the event was not triggered for manual constant assignments (think Hotel = Class.new). So, while that was a corner case, unfortunately we could not achieve full parity with this technique. The most important and common way to the define namespaces, using the class/module keywords, worked. But argghhhh. But the suffering is over my friends! In part motivated by this problem, @_byroot implemented Module#const_added. That callback is invoked in both situations! 🤘 Thanks to Module#const_added, the limitation has been finally removed. Also, Zeitwerk no longer uses TracePoint internally (/cc @yukihiro_matz). Zeitwerk 2.7 requires Ruby 3.2 or later, and upgrading should be seamless for existing projects. Don't worry if your app uses an older Ruby. Bundler realizes 2.7's Ruby version constraint is not satisfied, and ignores it in that case. Gemfile.lock will always get a version compatible with your app. Same for gems that need to support Ruby versions older than 3.2, a version constraint for zeitwerk like ~> 2.6 will do the job. Client code will automatically use a compatible one. Mission complete! 🎉

@ltb1963 I don't think it is really getting any coverage. I only just found out about the strike when I had a service outage caused by a cut wire, and there not being any technicians (rightfully) not available. Are any of the competitors really going to be any better?

Time to question your interest in using AT&T. We are in 4 th week of its employees out on strike.. Here is a company with 29 billion in Revenues asking their employees to pay half their own health insurance and offering a 2.9 cost of living increase and cuts in pension and cut

We're hiring engineer #4 at @scholarlysoft on-site in Seattle. If that's you, let's chat! DMs are open. scholarly.breezy.hr/p/a8959b02b8bd…

@nateberkopec Interestingly, this means that it works well for things that aren’t rapidly changing. For example: shell scripts 😅 I just wouldn’t recommend running them without a review first tho

Part of the problem with using ChatGPT for code is that it basically generates the code a junior dev from 4 years ago would have written. In the frontend world in particular, the community "best practice" has moved on wildly since the model was trained.

Very exited and proud to be sponsoring #railsconf this year. I’ll be at the open source hackday working on singed, a performance debugging tool! medium.com/gusto-engineer…

@mbhnyc @nateberkopec @GustoHQ The good news is that bootsnap can actually help optimize this when you use YAML.load_file!

@nateberkopec @GustoHQ Stop loading a zip code location yaml! 😂

.@GustoHQ's multi-million-line Rails monolith boots in 10 seconds:

@fxn @_beauraF @_byroot @shioyama @nateberkopec @GustoHQ Actually belay that, you can use it instead of require. I guess I'm asking if there's any better you patterns you know to register this kind of thing without having to load the classes?

@fxn @_beauraF @_byroot @shioyama @nateberkopec @GustoHQ How would you use this in the gist? I have seen this pattern of using a superclass keeping a list of known subclasses, and that requires loading the subclasses at some point