-
Stop Using next in Ruby Loops
LLM-generated Ruby code tends to reach for
nextinside.eachloops. It works, but it’s not idiomatic. When you find yourself usingnext, there’s almost always a more expressive alternative using Enumerable methods.A Simple Example
Consider this loop that collects names of active users:
names = [] users.each do |user| next if user.inactive? names << user.name endThe
next ifis doing filtering, but it doesn’t say filtering. It says “skip this one,” which forces the reader to mentally invert the condition to understand what’s being kept. Compare:names = users.reject(&:inactive?).map(&:name)The intent is explicit: reject inactive users, then map to names. Each method communicates exactly one operation.
Discrete Steps
Enumerable methods let you express collection processing as discrete, named steps:
select,reject,map,filter_map. Each step communicates intent. When you usenextinside.each, you combine filtering and transformation into one opaque block. The reader has to mentally simulate the loop to understand what it does.Named methods are also composable. You can add, remove, or reorder steps without restructuring the entire loop. A chain of Enumerable methods reads as a description of what the code does, not how it iterates.
A More Complex Example
Here’s a loop that processes orders with multiple conditions:
summaries = [] orders.each do |order| next if order.cancelled? next if order.total < 100 next unless order.region == "US" summaries << { id: order.id, total: order.total, customer: order.customer_name } endThree
nextstatements, each encoding a different filter. You have to read all of them to understand which orders survive. Compare:summaries = orders .reject(&:cancelled?) .select { |o| o.total >= 100 } .select { |o| o.region == "US" } .map { |o| { id: o.id, total: o.total, customer: o.customer_name } }Each step is readable in isolation. The pipeline reads as a description: reject cancelled orders, keep those above $100 in the US region, then transform to summaries.
One Operation Per Step
Enumerable has methods like
filter_mapthat combine more than one operation. They can be useful, but at the cost of obscuring intent. Consider:results = items .select(&:valid?) .filter_map { |item| item.compute_value }filter_mapmaps and removesnilresults in one pass. But the reader has to know that to understand what’s happening. Compare:results = items .select(&:valid?) .map { |item| item.compute_value } .compactEach step does one thing: select valid items, transform them, remove nils. The tradeoff is worth considering –
filter_mapis more concise, but.map { ... }.compactmakes every operation visible.What About Performance?
The obvious objection: chaining creates intermediate arrays, while the
.eachloop iterates only once. In practice, this rarely matters. But when it does, that’s whatlazyis for:summaries = orders .lazy .reject(&:cancelled?) .select { |o| o.total >= 100 } .select { |o| o.region == "US" } .map { |o| { id: o.id, total: o.total, customer: o.customer_name } } .to_aSame expressive chain, single-pass iteration. You get clarity and performance.
Conclusion
When you see
nextin a loop, treat it as a signal. There’s probably an Enumerable method that says what you mean more clearly. -
The REPL: Issue 138 - February 2026
GitButler CLI Is Really Good
GitButler is new to me. At first glance, the ergonomics look a lot like
jj. I don’t feel the pain as much withgit. Sure, I work on different branches all the time, but switching is not a problem, and neither is cherry-picking. Stacked PRs are more painful, butgit rebase --update-refsgoes a long way. It still requires multiple pushes, but I don’t do that often. Worth watching.Demystifying evals for AI agents
Anthropic published a deep dive into evaluations for AI agents and how to maintain quality. The research they share publicly is impressive.
-
I've Disallowed LLMs
As an experiment, I changed the
robots.txton this blog to disallow LLMs from crawling it. I saw a reference – lost now in my notes – to something like this:# Disallow AI training # OpenAI - https://platform.openai.com/docs/bots User-agent: GPTBot Disallow: / # Anthropic - https://support.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler User-agent: ClaudeBot User-agent: anthropic-ai Disallow: / # Google - https://developers.google.com/crawling/docs/crawlers-fetchers/google-common-crawlers#google-extended User-agent: Google-Extended Disallow: / # Apple - https://support.apple.com/en-us/119829 User-agent: Applebot-Extended Disallow: / # Common Crawl - https://commoncrawl.org/ccbot User-agent: CCBot Disallow: / # Meta/Facebook - https://developers.facebook.com/docs/sharing/webmasters/web-crawlers User-agent: Meta-ExternalAgent User-agent: FacebookBot Disallow: / # ByteDance/TikTok User-agent: Bytespider Disallow: / # Amazon - https://developer.amazon.com/amazonbot User-agent: Amazonbot Disallow: / # Ai2 - https://allenai.org/crawler User-agent: AI2Bot Disallow: / # Bright Data - https://brightdata.com/brightbot User-agent: Brightbot 1.0 Disallow: / # Huawei/PetalBot User-agent: PetalBot Disallow: / # Meltwater/YaK User-agent: YaK Disallow: / # Yandex User-agent: YandexAdditional User-agent: YandexAdditionalBot Disallow: / # A draft standards-based way to disable AI training. User-Agent: * # https://www.ietf.org/archive/id/draft-ietf-aipref-attach-04.html # Content-Usage: train-ai=n Allow: / # Last updated: January 10, 2026That change went live on 2026-01-13. Now, after roughly a month, I can see what effect it had on traffic, as measured in my server logs rather than JS artifacts injected on page load.
Can you spot when the traffic dropped off? Neither can I. 🤷🏻♂️
-
The REPL: Issue 137 - January 2026
AI is a horse
This is very clever!
Karpathy on Claude
AI capabilities are moving fast. Its hard to keep up and develop personal workflows. I find it useful to read what others are doing and take what fits me.
Taming Claude Code
In the same vein as above. I’m experimenting with creating skills after reading.
-
I've been POSSE all along
I recently learned that Publish (on your) Own Site, Syndicate Elsewhere POSSE is a thing – about 10 years old. It caught me by surprise. Not because I find it novel, but because I’ve always followed these principles.
In 1995 I learned about the web. Connecting to the internet meant using the same phone line for calls. You connected on-demand, when you wanted to check email or browse. Most people got their email address from their ISP. I realized that switching ISPs would mean changing my email. So I got an address I thought would last forever:
ylan@email.com. People didn’t believe me when I told them that was actually my address! I forget what happened toemail.com, and the domain now sits unused (it redirects tocompany.ai). But at some point, the terms of service didn’t agree with me. Being technically minded, I realized the only way to avoid churning my email address was to buy my own domain. I could (and have) changed providers, but my email would stay the same. At the close of last century I boughtsegal-family.com, and have used it ever since.In 2021, I decided to start this blog. When looking for hosting providers, CMSs, and static site generators, I read some advice: Whatever you pick, make sure you own your content. Don’t let a third party own it who might change their terms later. Given my email address experience, this seemed obvious. Fool me once, shame on you. Fool me twice, shame on me.
Don’t get me wrong – I’m not putting anyone down for reviving an older idea. I’m happy that POSSE has been gathering steam. It’s better for the web. I look forward to a more interconnected future. And if it helps to have a catchy acronym, that’s great too.