# Nate Berkopec The blog of Nate Berkopec, Ruby on Rails programmer. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Thoughts on LLMs in 2026 DevFeed: [Thoughts on LLMs in 2026]() Original publisher: [Read original article]() Published: 2026-05-23T00:00:00Z Content type: opinion Language: en Sources: [Nate Berkopec]() Topics: [Large Language Model](), [Artificial Intelligence](), [LLM evaluation / benchmarking]() Tags: [ai](), [language-models](), [llms](), [thoughts]() ### AI overview An opinionated assessment of LLMs in 2026 argues that they are not general artificial intelligence but increasingly useful tools for verifiable, agent-assisted software work. It describes progress as uneven and mostly steady rather than evidence of an imminent singularity, emphasizing persistent weaknesses in visual understanding and simple reasoning despite stronger language and mathematics capabilities. ### Source excerpt In May 2026, we are now five months post-Opus-4.5 and the great "Christmas Break Revolution" that saw us all hunkered down in front of our laptops over the new year. I'm now about nine to twelve months into not writing code by hand anymore. I more or less spend 100% of my time that I used to spend typing in IDEs instead talking to LLMs in Pi.11 Seriously, have you tried Pi yet? You're a hacker, aren't you? Why are you eating at the sloptrough of a trillion dollar company and whatever their product managers think you need in an LLM harness, and instead build your own like the hacker you were meant to be? Has the great Singularity come to pass yet? No. There's literally only one benchmark graph of LLMs which shows exponential progress22 , so of course that's the one everyone looks at. However, the rest of the synthetic benchmarks, and my own subjective experience, feel more like a steady, linear pace of progress. Artificial "intelligence" continues to be the wrong lens through which to view this technology we call large language models. "AI" gets us tripped up by a vision of a kind of "superset" of human intellectual capability, an artificial entity which meets or exceeds us on all dimensions. But LLMs and agents aren't like that at all, and they're not becoming less like that either. Their jagged "intelligence", as far as it can be said to be intelligence, simply gets more jagged over time as this radar chart of capabilities looks more and more ragged. LLMs can solve Erdős problems but not simple brain teasers about car washes, et cetera, et cetera. Their visual intelligence and comprehension in particular continue to be much worse than their command of language and mathematics. This word--intelligence--is leading people into fun science fiction thought experiments and useless "the singularity is around the corner!" drivel, and away from what LLMs are actually really useful for.33 The AI bigwigs essentially treated effective altruists as useful fools who kept saying "A ## Why 37signals' Code Looks Different: Small Teams, Narrow Scope, and Selective Hiring DevFeed: [Why 37signals' Code Looks Different: Small Teams, Narrow Scope, and Selective Hiring]() Original publisher: [Read original article]() Published: 2026-01-27T00:00:00Z Content type: opinion Language: en Sources: [Nate Berkopec]() Topics: [Rails](), [Code](), [Software as a service]() Tags: [code](), [open-source](), [rails](), [saas](), [team-building]() ### AI overview The article argues that 37signals' clean-looking Rails code reflects a distinct engineering strategy rather than inherently superior developers. That strategy emphasizes small teams, aggressively limited product scope, and hiring highly capable engineers. ### Source excerpt A recent episode of the Rails Business Podcast was about "striving for ideal code". The hosts are both typical "Rails Indie" owner/operators: a small team building a small SaaS. They were discussing a new podcast which is out from 37signals called "Recordables". The discussion was envy for how good and clean 37signals' code was. This is a fair assessment. They've been open-sourcing a lot of real world products lately, including: Campfire, the chat app. Fizzy, the Kanban tool. If you read them, the code looks great. We've only seen glimpses of this Rails style, which David has alluded to or shown small parts of in interviews or conference talks. Until now, we couldn't see what the Rails "house style" is because we never had a real open source application written in that style. Now we do. The discussion on the podcast was mostly a sense of envy of how good the good was, and wondering: "Why doesn't our code look that good?" Covet not thy neighbor's codebase. The reason why your code doesn't look "as good" as 37signals' is because your engineering strategy is not the same as theirs. 37signals, for all of its faults, has a unique engineering strategy, which I would summarize as: Stay small in terms of headcount to revenue ratio. Ruthlessly cut scope. 37signals' products usually do far, far less than their competition. Hire the top 10% of engineers. Most of the businesses I've ever worked for cannot say that any of these three points are honestly part of their strategy. Every time 37signals puts out a codebase or talks about their size in public, I am always shocked at how few LOC they are. Basecamp Next/2 was originally 10,000 LOC. Basecamp 3 was 18,000 LOC when released. Fizzy is 7,500 non-test LOC. Campfire is just 2,500 non-test LOC. That is fucking insane. 37signals employs 25-30 technical employees, meaning that for each individual codebase, they're maintaining under 2,500 lines per person. Most companies manage 10x that per headcount. Most indies ship 100,000+ line ## What Does It Mean to Be an Open-Source Project Maintainer? DevFeed: [What Does It Mean to Be an Open-Source Project Maintainer?]() Original publisher: [Read original article]() Published: 2019-10-07T00:00:00Z Content type: opinion Language: en Sources: [Nate Berkopec]() Topics: [Maintainers](), [Open Source](), [Code review](), [Code](), [releases]() Tags: [bug](), [code-review](), [documentation](), [maintainers](), [maintenance](), [open-source](), [pull-requests](), [refactoring](), [releases]() ### AI overview The article examines what it means to maintain an open-source project. It argues that maintainership includes responsibilities such as coding, documentation, issue triage, code review, releases, quality control, project direction, and encouraging contributions, while only some responsibilities require commit or package-publishing privileges. ### Source excerpt There are many different responsibilities for the maintenance of an open source project, and while some of them get called "maintainership", I'm not sure they should be. Let's zoom out for a moment. What labor must be performed, by someone, anyone - in any successful open source project? Writing code - fixing bugs, writing new features, refactoring old code. Writing documentation - explaining why things are the way they are. Recording intention and design, and how to use the APIs provided. Issue triage. Making bug reporters create reproducing examples. Closing issues that are stale or invalid. Providing feedback on feature requests. Code review. Looking at pull requests and providing feedback. Regular releases. Packaging up a release and pushing to the package namespace. "Green Button" control. Merging pull requests. Deciding what should be merged and what should not. This "quality control": is this PR good? Does it fit the goals of the project? Did someone review it? Vision/direction. What's our release schedule? What features will we not include? What approaches will we never take? Enlisting help. Encouraging more contributions and engaging more volunteers to do the above. What's interesting is that only a few of these responsibilities require what most people consider to be the mark of the maintainer: commit bit and rights to the package namespace on the appropriate package manager. In fact, it's really just one of the responsibilities that needs that all-important commit bit access ("Green Button" control), and pushing to the package repository needs a separate privilege. Everything else needs no privileges whatsoever. Unfortunately, the word "maintainer" has become something of a byword for a mythical cowboy-coder heroine, someone who spends 40 hours a week on their project writing new features, fixing reported bugs, spinning up VMs on cloud services to try to reproduce the more esoteric reports. As an maintainer of a popular OSS project, many people certainly ## How To Get A Computer Science Degree in a Warzone DevFeed: [How To Get A Computer Science Degree in a Warzone]() Original publisher: [Read original article]() Published: 2017-08-10T07:00:00Z Content type: opinion Language: en Sources: [Nate Berkopec]() Topics: [Computer science](), [Ruby on Rails](), [Ruby](), [Open Source]() Tags: [computer-science](), [open-source](), [ruby](), [ruby-on-rails]() ### AI overview This personal account follows Mohammed, a Syrian computer science student at the University of Aleppo who corresponded with the author about Ruby and open-source projects while living through the Syrian Civil War. After fleeing Syria for Turkey, Mohammed shares his story and the challenges of continuing his education and interests during war. ### Source excerpt In June of 2016, I received an email: The original email Dear sir: I'm from Syria , expected to graduate this semester (Computer Science) , and as you sir my role model in Ruby also Rails , I need some advice about open source projects in Ruby to learn from. Thanks in advance and sorry if I waste your time Actually, he wasn't just in Syria. He was at the University of Aleppo. His name was Mohammed. Russian drone footage of Aleppo in February 2016. Mohammed: 'The damage now is much much bigger than this video.' Youtube At this point, the Syrian Civil War had been raging in Aleppo for almost four years. One tenth of the total deaths of the war so far have happened in Aleppo. When Mohammed wrote this email in 2016, the tide was just starting to turn in favor of Assad's forces. You would be forgiven if, like me, you thought that life did not go on inside of a warzone, especially in a city where entire city blocks had been leveled to dust. But, it does - although life is hard, life goes on. Aftermath of the 2013 Aleppo University bombing. Mohammed: 'When I saw the photo of the University bombing in the post I saw everything: I saw who did it, how he did it, the bodies, legs, arms, blood, glass, melted metal, I was there. My heart beat is getting faster and the inhales and exhales is disturbing my body better to close this thing.' NYTimes The veil of ignorance had been stripped away. Here was someone my age, on the other side of the world, with the same interests as I do, but living in an actual warzone. He played video games with his friends, like I do, enjoyed programming in Ruby on Rails, like I do, but his university was being bombed. Mohammed and I continued corresponding. He never asked for anything except advice, and that was all I gave (or really, could give). In January of this year, Mohammed fled Syria and made it to Turkey, where he is officially a Syrian refugee. Mohammed with his public key fingerprint. We both signed this post, see the end. Keybase I'll turn ## How I Made $220,033.50 Self-Publishing a Book About Ruby on Rails DevFeed: [How I Made $220,033.50 Self-Publishing a Book About Ruby on Rails]() Original publisher: [Read original article]() Published: 2017-03-10T07:00:00Z Content type: article Language: en Sources: [Nate Berkopec]() Topics: [Ruby on Rails](), [Rails](), [Publishing](), [Programming]() Tags: [developers](), [freelance-career](), [publishing](), [rails](), [revenue](), [ruby]() ### AI overview A step-by-step account of creating and self-publishing The Complete Guide to Rails Performance, which the author says sold just over 1,800 copies and generated $220,033.50 in gross revenue. The article also describes how the product affected his freelancing career and lifestyle. ### Source excerpt Did I mention how awesome it is to live 30 minutes from a ski lift? In March of 2016, I launched a course called The Complete Guide to Rails Performance. Since then, I have distributed just over 1800 copies, for a gross revenue of $220,033.50 (about $1000/week). Although I launched the course in March, I had worked on it for about 4 months. Releasing the course has been a lifechanging event for me. For the first time ever, I made more money over a year from product revenue rather than service revenue. Releasing the course nearly doubled my usual freelancing rate, allowed me to change my lifestyle by moving to a ski resort town in New Mexico, and turned me into something of a minor Thought Leader (tm) in the Ruby on Rails field. In this post, I'd like to share my process for creating The Complete Guide to Rails Performance (henceforth "the CGRP"), what I'd do differently in the future, and how creating a programming information product radically changed my freelance career. It would be immensely helpful to me if other programming authors (self-published or not) would email me to share their own experiences and results. I'm not sure if I'm an outlier or just the average case! The Path to Making Money While You Sleep I ain't programmin' here no mo'! I was on the first season of ABC's Shark Tank when I was only 18 years old. I think "product money" is the eventual goal for nearly all freelancers or solopreneurs. There's probably a lot of developers sitting on the sidelines too wishing they could tell their boss to take this scrum and shove it. And, now that I've made some product money, I can tell you it is pretty awesome. Waking up with more money than you went to bed with is a very, very good feeling. I've been trying to "start a business" since 2008 (you may know me from my appearance on ABC's reality show Shark Tank in 2009). It's always been a struggle for me to find the magical intersection in the theoretical Venn diagram of my interests/skills and what people wil