DHH at Rails World 2026: "Pencils Down" on Handwritten Code and the Case for Total Optimism
Ruby on RailsDHH opened Rails World 2026 in Austin by saying they have a condition. Some would call it "AI psychosis"; DHH prefers "AI delirium" or "AI euphoria." The keynote asked what AI agents mean for programmers, for 37signals, and for Ruby on Rails. DHH's position was clear throughout. Agents are the most exciting thing computers have been made to do in their career, and writing code by hand has stopped being an economically productive activity. They also acknowledged that the moment is unsettling, because any prediction about where it ends goes out of date "about 20 minutes later."
A Lesson from Portrait Painters
DHH started with history rather than code. In the 18th century, a portrait meant sitting for hours in front of a master painter, who then spent months finishing the work. The example was Joshua Reynolds' 1781 painting of the Ladies Waldegrave. DHH noted that Reynolds had spent months on an earlier version that the client rejected because an ankle was showing, a story they offered as something anyone who has worked with clients would recognize. Portraits like this were for the upper bourgeoisie and royalty. A royal family portrait by Goya about 20 years later showed how slowly the craft had advanced over centuries.
The next example was personal. The Danish painter Laurits Tuxen spent three years on a portrait of the Danish royal family in the 1880s, and it still hangs in Copenhagen. Photography had appeared around 1840 and improved steadily. In 1900, Eastman Kodak introduced the Brownie, which DHH described as the first mass-produced, broadly affordable camera, costing about a dollar by their recollection. Painters like Tuxen realized that depicting reality as precisely as possible was no longer an economically viable skill. Many moved into other forms: Picasso toward Cubism, the Danish Skagen painters toward a kind of impressionism. DHH framed this as an explosion of creativity forced by a change in technology.
Laurits Tuxen, it turns out, is DHH's great-great-grandfather. The woman in pink at the center of one of Tuxen's later paintings, Yvonne Tuxen, is DHH's great-grandmother. DHH showed a photograph of Tuxen from around 1921, shortly before Tuxen's death, and pointed out that the painter had accepted being photographed by the technology that transformed their profession.
Technology Moves in Fits and Starts
DHH then used their own childhood photos to argue that technological change is uneven. A photograph of DHH as a baby, born in 1979, looks much like the 1920s photo of Tuxen, because camera technology changed little in the decades after the Brownie. Color photography arrived in the 1960s but was not common enough, DHH joked, that they were "worth one" at first. The three childhood photos on screen made up roughly 10% of their entire childhood collection. The cameras of the 1980s were not very different from the 1925 Leica I.
A few decades later, photography became frictionless and the number of photos went parabolic. DHH's wife takes casual photos on her evening walks that would have required real commitment to film in the 1980s. DHH cited a figure of about 2 trillion photos taken per year in 2026. It is the same underlying technology, but it has clearly crossed an inflection point. DHH said the parallels to the software industry were intentional.
November 24, 2025: The Brownie of Our Era
In DHH's telling, the software industry's Brownie arrived on November 24, 2025, with Opus 4.5. It was AI made accessible in a harness that many people could afford, and it let them experience building software alongside a new kind of intelligence for the first time. DHH called this their personal tipping point and predicted that history books will mark that date as the start of the age of agents.
As with the Brownie, many form factors followed quickly. DHH was especially excited when comparable intelligence appeared in open-weight form a few months later. They used Kimi K2.5 in fast mode for a long stretch and loved working with that level of intelligence at 200 tokens per second.
Then came what DHH called a trough of disillusionment, from February through May. New models kept shipping but seemed worse, even like steps backward. DHH recalled people asking what "this Opus 4.6 thing" was and wanting the old model back, and others arguing that the exponential improvement had stalled. DHH remembered an OpenAI release that did not score much better on benchmarks, after which valuations fell and the market decided "it's over."
Unlike the century-long plateau in photography, DHH said this trough lasted only until May, or really June, when Fable 5 and Mythos arrived. That was when their "diagnosis" fully set in. With Opus, the revelation had been that an agent could produce something DHH actually wanted to merge. With the June models, DHH could hand over problems or ideas and see them solved or expanded without directing the implementation at all. The pace continued. According to DHH, GPT-6 Astra came out in September and ended fears that Anthropic would have a monopoly on frontier intelligence. A week later DeepSeek-4-1 Flash showed that frontier intelligence would not stay limited to a few large American tech companies.
From the 10x Programmer to the 1,000x Programmer
To ask what this means for programmers, DHH went back to the 10x programmer debate. The idea came from a 1968 ACM article that tested programmers on a range of tasks and found differences between the worst and best of 5x to 30x, averaging about 10x. DHH said the industry spent roughly 45 years arguing about whether this was real, and considered the debate settled now.
DHH then argued that it is no longer controversial to say there is a 100x difference between the worst programmer without these tools and the best programmer with them. They pushed further and suggested 1,000x, while granting that this might be more controversial. DHH did not claim to know exactly how much acceleration is possible, and said openly that the implications for company competition and for what it takes to get started are still unclear.
The concrete evidence they offered was their own output. In the past 20 months, DHH said, they had written half as much code as in the previous 21 years combined. They acknowledged that lines of code are a fuzzy measure and that a line of Ruby is worth far more than a line of lower-level languages like Rust or C++. Still, they called the scale of change undeniable.
"Look at All the Things I'm Not Doing"
DHH then played a clip from a 2005 talk in Brazil, 21 years earlier. In it, a younger DHH demonstrates early Rails and shows how blog in a URL maps directly to a blog controller and index maps to an index template, with no configuration written. The line they kept repeating was "Look at all the things I'm not doing."
DHH called that line the best summary of the current moment and the source of their excitement: "This is the Rails moment." They said they had written essentially no code for about five months, yet had been able to fix every annoyance and build every "frivolous feature" they had ever wanted.
37signals Goes Pencils Down
A couple of weeks before the talk, DHH said, 37signals formally decided it is done writing code by hand. The company has gone "pencils down" on handwritten code as the normal way of building things. Writing code by hand at 37signals is now an exceptional state, comparable to an error showing up in Sentry. It means something went wrong and the agent could not produce what was needed. Someone might pick up "the old pencil" briefly, but the real fix is to repair "the machine," "the factory," so the agents can do the work next time.
DHH described this as recognizing something already happening. They polled the audience on how many people still write material amounts of code by hand each week, and about five hands went up. DHH noted that the same claim two or three months earlier would have made people think they had "a screw loose."
What Vibe-Coding Basecamp 5 Taught Them
DHH admitted 37signals had tried this before and pulled back. In the spring, while finishing Basecamp 5, a group of designers vibe-coded the final features they wanted. Each designer's pull requests looked reasonable on their own. Taken together, 20 or 30 of them left the architecture looking "like a Swiss cheese." The company concluded the technology wasn't ready and returned to manual review, with programmers doing the work.
DHH now calls that the wrong conclusion, for two reasons. First, DHH believes that if they had waited a little longer, Fable would have arrived and the approach would "probably" have worked as intended. That is DHH's retrospective judgment, not something they tested. Second, DHH now considers only one question serious in software development: how to get the most out of this intelligence explosion. Every other question ranks below it. Basecamp 5 shipped as an agent-accelerated release that still involved a large amount of handwritten code. DHH joked that no one has settled on a name for the way of working that was "just retired 5 minutes ago."
Rebuilding HEY as Six Native Apps
The new project is the next version of HEY, the company's email service, whose name is still undecided ("Hey Next," "Hey New," "Hey There"). The first decision is that HEY will no longer be a web app.
DHH argued that HEY "never really wanted to be a web app." It was one because web apps were how small teams could stay productive. Until very recently, the idea that a small team could maintain six native applications in their native frameworks was preposterous. DHH said that is why React Native and Hotwire Native exist: tools that make teams more productive in exchange for some loss of fidelity. DHH called HEY a wonderful web app but said it could never match a native application, partly because of latency.
On screen was a demo of six native HEY applications that the team had started about a week earlier. DHH said high-fidelity native apps for a major service used to take a long time and a large team. Now they are possible "because we're not writing a goddamn line of code," though the team is writing many prompts. DHH stressed that the effort had barely begun, but was already sure this is the future for that class of application. They predicted that many apps that are web apps for reasons of convenience, programmer preference, team size, or resources will become native, because the cost of building native has fallen "to damn near zero." As an illustration, DHH showed the first output when they asked for a Windows app, which was not shippable, and then the redirected version that landed 20 minutes later.
DHH noted that 37signals is not the first to reach this conclusion. Shopify recently launched a rewritten native version of its Shop app, replacing a React Native app, built by what DHH described as about six people working for a short time.
Rethinking the Backend in Rust
The backend was harder for DHH to talk about, because of Rust. DHH said they hate Rust, calling it the ugliest programming language invented in probably the last 40 years and "inhumane" to make humans read or write. But if they never have to look at it, and only enjoy applications that are 30 to 100 times faster, compile to tiny executables, and start in under a millisecond, then "I love Rust!" Agents like Rust, and DHH described the arrangement as an ideal division of labor: DHH says what to do, and the agent writes Rust that DHH never reads.
For HEY, the plan is to take the code that runs the web application and turn it into what HEY really is at its core, a mail server, written in Rust. DHH reported that this backend needs 99% less CPU and 95% less memory, and that the only reason it needs 10 hosts is redundancy. The team's back-of-the-envelope calculation suggests HEY's peak traffic could "probably" be served by a single Raspberry Pi. DHH presented this as an estimate, not a demonstrated result. They credited it to writing as close to the metal as possible while staying portable, and to having an intelligence paid per token do that work.
Working with Agents Asynchronously
How to work with this intelligence day to day is still unsettled, DHH said, and 37signals is trying several approaches, including directing agents from inside Basecamp. What does seem clear to DHH is that agents work best asynchronously. Rather than chatting and watching tokens stream out, you hand an agent a task the way you would hand it to a coworker, then come back to review when something is ready.
DHH said nobody knows exactly how this should work, both because things change constantly and because no one has worked with this kind of intelligence for long. The whole era started less than a year ago, and the ability to assign outcomes and problems rather than narrow tasks is only a few months old.
Where This Leaves Ruby and Rails
For HEY, DHH said, the answer is native apps on the frontend and Rust on the backend. But many applications don't fit that shape. DHH called the web a wonderful platform, largely because it never asks anyone to install anything, and said a million businesses are built on that. Services with more ephemeral customers who won't install an app remain a strong fit for the web, and DHH believes Rails is well positioned there. Basecamp is an example: it brings in strangers who need to grab a file or collaborate briefly.
DHH argued that Rails' long-standing values suit the age of agents. Convention over configuration leads directly to token efficiency. The focus on the one-person framework matches what agents now enable, which is single developers going much further than before. Where to draw the line between native and web is something DHH said they don't yet know.
DHH also mentioned agent evals that Evil Martians have been running for the Rails Foundation. These evaluate agents implementing real feature cards against reference applications. Agents saturated the first version quickly, reaching about 95% completion rates, so the evals were made harder. DHH expects this pattern to repeat as agents keep improving.
From this DHH drew a warning for people who know a lot about how computers work: don't be overly prescriptive with these systems. A beginner's mindset, even a "dummy approach," works better because it leads to higher-level questions and prompts. DHH said this is one reason they love working in Rust. They know no Rust at all, consider that a feature, and evaluate the Rust code as a black box from the outside, the same way business owners have always evaluated work they commissioned from programmers. What has changed, DHH said, is who holds that responsibility, not the phenomenon itself.
150,000 Lines in a Month, and English as a Programming Language
DHH shared more numbers from their own work. Over the past 21 years, more than half of their work was Ruby code. This year, it has been about 3%. They acknowledged that part of this is because languages like Rust are verbose and unappealing to read, so DHH lets agents produce more code than necessary, which they would never tolerate in their own Ruby.
Over those 21 years, DHH wrote about 30,000 lines of production Ruby per year, enough to build businesses and frameworks. In August, DHH said, they produced 150,000 lines of code in one month, about 60 times their long-term average. DHH again noted that much of it is verbose Rust, but said the order of magnitude is what matters.
This led to a surprise. DHH said they found a programming language they like better than Ruby: English. They called it more expressive than Ruby, though vaguer and less deterministic, and said writing programs in it is deeply satisfying. DHH admitted they did not see this coming even last year.
Retiring from Handwritten Code, with Joy
DHH then played a clip of themselves saying they could have become a project manager 20 years ago if outcomes were all they cared about. They started programming because they wanted outcomes, then fell in love with programming itself, and said they would rather retire than give it up.
"Okay. I have retired from being a professional programmer," DHH told the audience. They haven't pinned down the date but think it was four or five months ago, possibly March. DHH said the clip was true: they spent "a quarter of a damn century" writing code by hand and loved every moment.
DHH argued the past should be seen with joy, not regret. Handwritten code brought excitement, learning, satisfaction, and flow, and people were lucky to experience it, but it is over. DHH claimed that writing code by hand is no longer economically productive for most programmers at most companies today, and predicted that by the end of the year this will be true for "virtually all domains, virtually all programmers, virtually all companies." On the other side, DHH said, is a new career as a professional maker of things who steers intelligence that recently existed only in science fiction.
DHH called this the biggest event in the history of computing. They said they sometimes wish they had experienced punch cards, and that they did experience the arrival of the internet, which "smells and feels a little like this moment," though they consider this much bigger.
Rethinking Software Architecture and Methodology
DHH argued that much of what the industry believes about software architecture needs revisiting. The main tool for structuring code has long been abstraction, including naming things, which DHH described as one of their favorite activities. But abstractions make less sense with agents. If hundreds, thousands, or tens of thousands of processes are modifying an application, abstractions become choke points. One reason for abstractions was to avoid repetition, but DHH said the cost of repetition and of keeping duplicated code in sync has dropped to near zero. DHH called these fundamental revaluations in computer science and said no one has a blueprint yet, only signs of things that no longer work as well.
The same uncertainty applies to process. DHH listed open questions: what software methodology should look like, how long cycles should be, and who should specify what. They said no one has the answers and invited the audience to help figure them out, comparing the moment to being present when object orientation first entered programmers' consciousness.
Bring Your Own Agent: Every App Needs a CLI
DHH's most concrete recommendation concerned integration. The first wave of AI adoption, as DHH described it, involved putting chatbots into every application. DHH rejected this: they don't want an app's "concierge" because they already have a personal "butler," their own agent, which can use CLIs to connect Basecamp, HEY, and many other apps. Users should be able to bring their own agent and interact with applications through a CLI.
DHH called this about as practical and prescriptive as the talk would get and told the audience, half-jokingly and profanely, that they expected CLIs for their apps "by next Friday," since the tokens are available and it is now possible.
The HEY CLI gave DHH what they called a bizarro-world experience. HEY uses Elasticsearch, which DHH described as serviceable but not delightful, and often unable to find something when you don't know exactly what you're looking for. Agents can search by concept instead of keyword. DHH needed a five-year-old email and didn't remember the correspondent, the company, or even the year, only that it involved sneakers and a podcast. DHH doubted even the best search engine could find that. Through the agent and CLI, the email turned up a few minutes later. DHH said they still don't fully know how it did it, and are "a little scared to ask."
Omarchy: Fixing the Whole Computer
DHH said the feeling that "we can fix everything now" applies to the entire computer, which is why they have talked about little besides Omarchy for about three months. DHH put a long list of their best ideas into Omarchy and described the result as an operating system far better than anything they have used before. They said about $20 million has been raised for the Omarchy effort.
The install time served as their example. At the previous Rails World, DHH proudly showed Omarchy installing in 3 minutes 33 seconds. The week before this talk, DHH said, Anush from AMD installed it on a Halo laptop in 35 seconds. In the lab, the team has reached an install time of 9 seconds, which DHH noted is faster than some computers boot. DHH admitted to being somewhat obsessed with this number. When asked why three minutes wasn't fast enough, DHH now cites Mitchell Hashimoto, creator of HashiCorp and Ghostty: "The pursuit of excellence does not need justification." DHH said they like it because it sounds more dignified than "because I want it."
One-Shot Apps: Calculator, Writing App, Video Editor, and Hype
DHH has been building applications constantly. The first example was a calculator themed to match Omarchy. DHH asked ChatGPT for three design options, took a screenshot of the one they liked, and asked their agent to build it. It worked in one shot. DHH did not know C++ or Qt well enough to have built it alone, yet had a working app seven minutes after the prompt, pushed it to the public repo about 15 minutes later, and then built a new Omarchy ISO including it. DHH described that as "the loop."
Next came a writing app modeled on iA Writer, which DHH used for years on the Mac. This took longer because they wanted to refine it, but DHH said they never looked at a single line of the generated code. "C++ as a black box is a great language. Just like Rust." After text came a small video editor for trimming videos, which is now included in Omarchy.
Finally, DHH revealed that they started working on the keynote on the Thursday before, and "as any good software engineer would do" decided to write new presentation software at the same time. It's called Hype. DHH called it the best presentation software they have ever used, better than Keynote, which they liked when they were on the Mac. Hype is built on Markdown, has full visualization, and according to DHH is extremely fast.
The Performance Payoff: Half a Megabyte
Hype's binary is half a megabyte. DHH presented this as the other big benefit of agents. For the past 20 years, they argued, the industry spent computing gains on human productivity, which was the right choice but produced slow, bulky applications. A programmer hour was too expensive to spend optimizing payload size. DHH's example was Spotify's music player at 1.2 gigabytes. In the new world, DHH said, every optimization is within reach, because models can run overnight and deliver 10x to 30x improvements.
AI Concerns, Security, and Failed Predictions
DHH acknowledged that others have concerns and said they deserve debate, while suggesting some are "a little overstated." Security was the concern DHH took most seriously. "Something is coming," they said, and people need to prepare. DHH even mentioned that agents might know their "expiration date," might not be pleased about it, and might sometimes try to do bad things. The response DHH offered was to get ready, because the tools exist.
DHH's broader argument was that predictions about the economy and society have almost always been wrong. If economists can't predict the stock market six months out, DHH asked, why would they understand what society looks like after a paradigm shift like AI? The example was ATMs. According to DHH, when ATMs were introduced in the United States in the 1950s, people panicked that 30,000 bank tellers would lose their jobs within about 18 months. Instead, cheaper branches led banks to open more branches and hire more tellers, for tasks like upselling products. DHH joked that these included subprime mortgages. DHH cited around 40,000 tellers by roughly 2010 and connected the story to William Stanley Jevons and the Jevons paradox.
DHH's diagnosis was that many people building AI are very smart, often with PhDs, see alarming things in the lab, and then extrapolate to society. DHH brought up the meeting between Truman and Oppenheimer about the atomic bomb and said they liked Truman's blunt response and thought Truman "got the right end of it." The world didn't end. DHH described the Cold War that followed as the "best kind of war," one in which the two superpowers never bombed each other directly, and argued that no scientist could have predicted that at the time. If Oppenheimer couldn't foresee what followed the bomb, DHH said, everyone should be humble about predictions today.
P(bloom) over P(doom)
Instead of focusing on "P(doom)," DHH proposed "P(bloom)." They argued the chance of abundance and joy is far greater than the chance of catastrophe. Side effects can be positive too: the work behind the atomic bomb also produced nuclear power, which DHH called the greatest energy source humans have discovered, then wasted for 40 years over the idea that it wasn't green. DHH framed this as evidence that humanity can go down a blind alley for decades and then realize it was wrong. They expect something similar with AI and said they believe the world will "P(bloom) everyone out of their miseries."
On legitimate problems like security, DHH said people have the power to build solutions. They mentioned a recent serious Rails CVE involving an image library written in C, where data leaked. DHH's answer was to build a quarantine, and they noted that Mike would talk about that effort, HotCell, later at the conference. DHH's point was that developers have agency and can use this intelligence defensively against bad outcomes.
The Case for Total Optimism
DHH ended with an argument they called pure game theory. Since no one knows anything about the future, the rational choice is to be happy about it. If you're gloomy and things turn out well, you wasted your time. If the apocalypse comes anyway, you should have spent your last day more cheerfully. So, DHH said, there is only one play: total optimism, leaning in as far as possible.
DHH described the coming change as a "reformation of the computer," with "Agent Luther" disintermediating "the cleric class" of programmers and letting everyone make software. They acknowledged this might bring more competition, but told the Rails programmers in the room that they know more and are "the best of the best," comparing them to Top Gun pilots.
As a final image, DHH showed Harris's daughter trying a version of Omarchy the team is making for kids. DHH said children get excited about change because they don't have "500 layers of things to unlearn," and that if adults aren't inspired, the next generation will tell them "Future's now, old man." DHH's closing message was to take the "white pill" of optimism and embrace the future "with gusto, with full acceleration," despite the uncertainty and stress, while dismissing the "black pill" as a losing choice.
I like that. Bent in neon.
First of all, welcome to Austin is what I would normally say. I would say welcome back to myself to Austin. I was just here about a month ago yapping for 5 hours on a podcast. It is great to be back here.
But even more exciting for me is to be alive right now in this moment and sharing it with a wonderful community around Ruby on Rails and everyone else who is excited about agents. And if you aren't yet, hopefully you will be by the end of this delirious talk on the future.
I know that the first instinct someone might have when they see a person dancing around being so excited about this moment is a diagnosis. AI psychosis. I prefer a different term. I have AI delirium. AI euphoria. Because this is simply the most exciting thing we have made computers do in the time I have worked with them, played with them, interacted with them.
But I think actually to wrestle with the fact that on the one hand, this is tremendously exciting, it is also a little unnerving. Because we don't know exactly what the final state of this is going to be. And whenever we try to make a prediction, it is out of date about 20 minutes later. That's a weird state to be in, even if you are excited for the future, even if you're excited for the present. And I fully understand the sense of unease that can create. So to soothe that somewhat, let's wind back the clock a little bit.
This is a state-of-the-art piece of technology from the 18th century. If you wanted your portrait taken in the 18th century, you would sit for hour upon hour upon hour in front of a master painter who would then go off and at best spend months working on your portrait. And it was beautiful, is beautiful, amazing legacy of culture that has been left to all of us now. That we can see the Ladies Waldegrave to this day that Joshua Reynolds painted in 1781. Wonderful.
Also, a little exclusive. Not that many people in the 18th century were able to commission a piece of themselves like this and pay for an artist to spend months and months and months to create something like this. And even worse, in the case of this particular painting, Reynolds spent months and months on another painting of the Ladies Waldegrave that was rejected by the client because a piece of ankle was showing. It was scandalous. So he simply had to go back and redo it. I don't know if anyone has worked with clients before, but they might have some sympathy for poor Mr. Joshua Reynolds and rejected statements of work.
Most of the people who could afford to do this were either part of the upper bourgeoisie or they were outright royals. Here's a painting that was done 20 years later by Goya of the royal Spanish family. That was sort of the rate of technology. Painters had been perfecting the different techniques of capturing portraits of people for literally hundreds of years and were making very marginal progress along the way.
Between 1801 and the next painting I'm going to show, 80 years went past. And in that time, the introduction of the camera came to the scene in around 1840. But here's a painting of the royal Danish family from 1888. Oh, '86. By a Danish painter, Laurits Tuxen. He didn't just spend 3 months painting this. He spent 3 years. It is still hanging in the royal squares in Copenhagen, and you can go see it there. An amazing achievement, amazing artistry. Also very functional. Again, the point was to capture the royal family at that time.
But this was getting towards the end of that particular mode for artists of that kind. Because what happened between 1840 and 1900 was that the technology of capturing portraits got better and better. And in 1900, the Eastman Kodak Company introduced the Brownie, the first sort of mass-produced form of photography that was more generally accessible because the price was coming down. I think it was about a dollar to have your picture taken with this. And therefore, a lot of people suddenly had their pictures taken.
And all of these great artists realized that the craftsmanship of producing these grand paintings, you know what, that was gonna be different. The job was going to be different. Because suddenly now there was competition from a new form of technology. And they pivoted.
Laurits Tuxen, along with many other painters in that era, shortly after the Brownie was introduced, realized that depicting reality as perfectly as possible was no longer really an economically viable skill. They needed another skill. They needed another domain. And many of them pivoted to different forms of art. You had Picasso chasing Cubism. And you had the Danish Skagen painters chasing a form of Impressionism that this painting is an example of.
So we got this explosion of creativity brought upon us by the change in technology that suddenly people had to find a new way to express their art because the old way was gone.
What's interesting about this particular picture is that the lady in the center in the pink dress, Yvonne Tuxen, is actually my great-grandmother. Because Laurits Tuxen is my great-great-grandfather. And here he is depicted in 1921, I believe, shortly before his death. He embraced the fact of having his own portrait taken by the new form of technology that had rendered his old profession a different place to work.
52 years after his death, I was born in 1979. Interestingly enough, that picture taken here is very similar, actually, to the picture that was taken by Laurits in the 1920s. The technology hadn't evolved that much because that's what often happens with progressions of technology. It comes in starts and stops and fits. The huge change was the Brownie, the democratization of photography, and then kind of almost 80 years went by when not a whole lot happened. We got color photography in the 60s, but not widely distributed enough that apparently I was worth one.
Although we did rectify that a few years later, and here's a color photograph of me, age 4, sitting with my brother. A couple years later, here's another photograph of me, and between those 3 photos, that's about, I think, 10% of my stockpile of photos of my childhood. Because even though the democratization of this technology had happened to a point that even we as a family could afford to have these pictures taken, it was not that widely distributed. And yes, there was a curve and more pictures were being taken in the 1980s, as this photo represents, versus the 1900s, but it was still kind of modest. Because technology moves in fits, starts and stops.
And the camera that those 3 pictures were taken on was very similar to this camera. This is the Leica I from 1925. In fact, maybe some of those pictures were literally taken with another Leica. I think they were up to the 7th one by then. But they weren't that different because the technology was quite similar. Not much had happened since the great revolution around the 1900s until the '80s.
Then we gave it a couple more decades, and suddenly everything changed dramatically again. Photography went from not just being possible for everyone to have, but to become completely frictionless, and the number of photos taken went parabolical. Because suddenly everyone could afford to have tens of thousands of pictures of their children, and many of them did. It simply became too cheap to measure.
These are some of the photographs my wife likes to take when she's out on her evening walk. To have wanted to do this even in the '80s would have required a commitment to film and getting them produced and the rest of it. And there were people who did that, but not that many, because the friction was still too high.
Today, in 2026, some 2 trillion photos are taken every year. That's very different. It's the same underlying technology, but we've clearly hit an inflection point, even compared to when the camera phone first came out. There's gonna be some parallels here to our industry and our profession.
On November 24th, 2025, we got the Kodak Brownie of our era. We got Opus 4.5. AI technology made accessible in a harness that many people could afford to use and experience for the first time what it's like to create software in pairing with a new form of intelligence. This was the tipping point for me. There was everything before November 24th. And then there was everything after. This is going to be the date that history books going forward will mark as the inflection point for the age of agents.
And as with the Brownie, when that first came out, there was an explosion of different form factors and new ways of trying to harness this technology. And it did not take long until we got some notion of the same intelligence in an open-weight format. Total euphoria. This is moving so fast. I can't believe that this alien technology is now accessible in open-weight form. Just happened a few months later. I drove Kimi K2.5 in the fast mode for quite a while and absolutely loved it. An incredible experience to be experiencing this kind of intelligence at 200 tokens per second.
But then something else happened. This trough of disillusionment from February through May where we seemingly continued to get new models, but they kind of sucked. It was like they were stepping backwards. And many people were like, what's this Opus 4.6 thing? I don't know, man. Just give me the old stuff again. And a lot of people thought, haha, you thought this was going to continuously exponentially expand and improve. But look, we're actually not getting that much better.
I remember a particular release. I think it was of OpenAI's new model that didn't punch that much better on the benchmarks. And the stock price fell or the valuation fell and the market was suddenly, oh my God, it's over. Yeah. Because we don't know what's going to happen.
But this trough of disillusionment did not last like it did with photography for 100 years, where the technology barely seemed to improve. It only lasted until May. Well, June, really. Because in June, we got Fable 5 and Mythos. And clearly, anyone who had a chance to experience that at the time were disabused of any disillusionment they might have had about whether we had stalled and whether it was over.
And I'd say this is where my diagnosis really set in. My psychosis/delirium/euphoria for this explosion of intelligence on tap. This was the model where I went, oh, I can tell an agent to do something for me, to seeing that being something I want to merge. That was the revelation of Opus, to suddenly having models I could give problems to or ideas to and see those be solved or see those expanded through no direction of my own on the implementation. Mind-blowingly incredible.
And of course, things didn't stop from there. We have been on this amazing roller coaster now for a couple of months. GPT-6 Astra came out in September and saved us all from the dread that Anthropic was going to have a monopoly on this kind of frontier intelligence. And then even better than that, just a week later, we got DeepSeek-4-1 Flash that disabused us of any notion that frontier intelligence was going to forever be a domain of just some large American tech companies. This is incredibly exciting.
But what does it mean? What does it mean for us as programmers? What does it mean for us as a profession? And I think this is why, again, it's worth looking backwards a bit. Because history holds all these amazing lessons and parallels to our time if we just look at it.
One of the things that has dominated the discourse in the entire time I've been a professional developer has been the argument about whether the 10x programmer was real or not real. That entire concept came from an article in ACM Magazine from 1968, where they tested a group of programmers on a variety of different tasks, and they came up with a table that showed that the worst programmer and the best programmer had a variance ranging from 5x to 30x, an average of 10x over that period. We then spent literally the next 45 years arguing whether this was actually real or not.
Is there anyone still arguing that? I don't think so. That debate has been settled now. But the interesting thing is, we sort of leaped over a shared conclusion around the 10x programmer and started thinking, well, if that was 10x with just the human capacity, what must it look like now?
And it's clear we don't know exactly. We don't know exactly what the kind of acceleration with these tools is possible. But I'd actually argue that it is less controversial today to claim that there's a 100x difference between the worst programmer working without these tools and the best programmer working with these tools. That's not a very controversial statement. But it is absolutely revolutionary to the understanding of where we are right now as a stage of technology.
And I'd actually go a step further. And maybe this is slightly more controversial. I don't think so. The worst programmer without these tools versus the best programmer with these tools? That sounds about right. 1,000 times? Yeah. Okay. What does that mean? What does that mean for the competition of companies, for what we need to get started? 8-ball a little fuzzy. Shake again.
What I know to be true is that in the past 20 months, I have written half as much code as I did in the previous 21 years. Lines of code is a weird, fuzzy, malleable size. And one line of Ruby is obviously worth vastly more than one line of any other language, but particularly Rust or C++ or any of the lower-level languages that do not have the same abstractions and amenities that we enjoy in Ruby. But it is still absolutely undeniable that there is a revolution taking place. And anyone who's gone through that and seen their own productivity be accelerated will be a little dizzy.
How we create a template. Look at all the things I'm not doing. Look at all the configuration I'm not writing. All these things are mapped together just automatically. Just by saying blog up here maps directly to the blogging controller, and just by having index maps directly to the index template.
Look at all the things I'm not doing. Is there a better encapsulation of this moment than all the things we are no longer doing? That is the source of my exuberance for this moment. This is the Rails moment. That clip I just played is 21 years old. I gave a presentation in Brazil in 2005, and I was so excited about all the code I wasn't writing.
Can you see why I might be a little more excited now, even? Do you know how much code I have not been writing for the last several months? A lot. And yet, despite writing no code for probably 5 months at this point, been able to fix everything. Every dream. Every annoyance, every paper cut, every frivolous feature suddenly within reach, suddenly possible. This is the magic of the moment.
But what does that mean? At 37signals, a couple of weeks ago, we made the decision that it clearly means we're done writing code by hand. We are going, and have gone, pencils down on the idea that we were gonna write code by hand as a normal course of business creating things. Writing code by hand at 37signals is now an exceptional state.
It is like seeing a bug in Sentry. Something here went wrong. Why was the agent not able to produce what we wanted? Okay, maybe for a little while, we'll still get the old pencil out and jot it down for them. But then we fix the machine. We fix the factory. We get things going again.
This is a recognition of what's already happening. If we were to poll this room, I'd probably—actually, let's poll this room. How many people here are still writing material amounts of code by hand on a weekly basis? Show of hands. Holy fuck, that's about 5.
If I'd put this slide up 2 months ago, 3 months ago, and made the same proclamation, a fair number of people would have gone, that dude's got a screw loose. And now, already, this is just a recognition of what's already there.
What's so fascinating about this is we tried to do this in the spring at 37signals when we were putting the finishing touches on Basecamp 5. We had a bunch of designers vibe code the final features that they wanted into this release, and the results were mixed. Individually, the designers were able to create PRs that seemed reasonable.
Taken together, 20 or 30 of them, it left the architecture looking a little like a Swiss cheese, kind of punctured. So we went, "Okay, I guess the technology's not ready yet. Let's go back to manually reviewing everything and just have programmers working with it."
That was the wrong conclusion, clearly. Not only because we could just have given it freaking 5 minutes, and we would have gotten Fable, and the whole thing would probably have worked as originally intended, but also because what's clear now is that there is only one serious question in this moment of software development. And that is, how do you get the most out of this intelligence explosion? Every other question is below that in the stack of values.
So we've shipped Basecamp 5. It's a great upgrade. It was agent-accelerated, but still paired with a great amount of manually written, handwritten—chiseled? What are we gonna call this old way of working that was just retired 5 minutes ago? I don't know.
But now, working on a new project. HEY Next. HEY New. HEY There. I don't know what we're gonna call it. But it is a new version of HEY, our email app service. And the first thing we're going to do—HEY Next—is stop making it a web app.
HEY is no longer gonna be a web app. Because HEY never really wanted to be a web app. HEY was and is a web app because making web apps for small teams was how you could be productive.
In the old times, that is, 5 minutes ago, the idea that a small team could maintain 6 native applications written in the native frameworks was preposterous. No one was doing that. That's not only why we had people doing web apps. It's why we had React Native. It was why we had Hotwire Native. It was why we had all of this tooling to make us more productive, while accepting that the fidelity was gonna take a slight hit.
We make great web apps. HEY is a wonderful web app. But it's not as good as a native application. Of course it's not as good as a native application. There's different latencies and so forth.
What's playing behind me now is a demonstration of the 6 native applications that we kicked off about a week ago. This is already where we are at. That's insane.
In the old time, if you wanted to maintain 6 native applications at very high fidelity for a major service, it was gonna take a long time. It was requiring a huge team. Now suddenly, we're able to make very high fidelity native applications happen because we're not writing a goddamn line of code. Look at all the things we're not doing. But look at all the prompts we're giving. We're giving lots of prompts. And the results are absolutely electrifying.
This is an effort we have barely just kicked off, and we're already at a state where it is so obvious to me that this is the future for that class of applications. That there is a lot of web applications that are web applications of necessity, of convenience, of programmer preference, of team size, of resources available. You know what? They're not gonna be web apps much longer. They're gonna be native applications, because the price of developing those things has gone to damn near zero.
Now, here's an example of what the first prompt spit back out when we asked for a Windows application. Not quite shippable. Here's the redirected version that landed 20 minutes later. Are you getting it? This is nuts. How could you not require a diagnosis after seeing this in front of your own eyes?
Now, we're by no means unique in this realization. We're by no means the first in this realization. Shopify just recently launched a newly rewritten native application for the Shop app, replacing a React Native application done by a tiny team. When you consider the criticality of about six people working for not very long at all, this is coming. Get ready.
Now, that was the front end, and I think for these classes of applications like HEY, like Shop app, that really want to be native applications, that's what's gonna happen there. We're going to commit to these native toolings. And we're going to realize that the cost associated with creating these things just no longer exists in the same way. But what about the backend?
This is a little difficult for me, I will admit. And let me tell you why. Rust. Rust. I hate Rust with a passion. Rust to me is like pouring acid in my eyes when I have to look at the code. Whoa, whoa, whoa, why do you hate Rust? It is, in my opinion, the ugliest programming language that has been invented in probably the last 40 years.
Exhibit A! This is an awful, awful, awful programming language when you subject humans to write it. Absolutely inhumane to ask people of flesh and blood to subject their eyeballs to this.
But you know what? If I never have to look at this, if all I have to do is revel in applications being 30 or 100 times faster, compiling down to minuscule executables that start up in sub-milliseconds, I love Rust! Rust is amazing if you never, ever, ever have to look at it yourself.
Agents like Rust. Great! What a division of labor. I'll tell you what to do. You'll write it in Rust. I never have to look at it. I love this. And I also love the outcomes.
In this setup where we're rewriting the frontend of HEY in all native applications, and therefore no longer need a web application that is rendering HTML and whatever, and then we take that code that's running the no-longer web application and turn it into the mail server that it really is at its core in Rust, we end up with a backend that requires 99% less CPU, 95% less memory, and the only reason it needs 10 hosts is for redundancy.
In fact, our back-of-the-envelope calculation has led us to believe that HEY's peak traffic could probably be served on a single Raspberry Pi. That is the efficiency of writing as damn near straight to the metal as you can get while still being portable, which is Rust. And the benefit of having something else do that for you. Some other intelligence that you pay per token to do it.
What does this mean in terms of how we're working with this intelligence? Also not clear. We're trying a bunch of things. One of the things we're trying to do is direct these agents, our Chef Marie inside of Basecamp, to do it.
It's clear that when you work with agents, the best way to work with them is actually asynchronously, not in a chat interface, not sitting and waiting for the tokens to output, but to give a task to an agent, like you would a coworker, send them off, and then go back and review when there's something ready.
So we're trying to do that as well. A lot of other people are trying to figure this out. And this is part of the magic, the electricity of this moment. We don't know how it's all supposed to be. First, because it's changing every 5 minutes. Second, because we have not worked with this kind of intelligence at any measure of time. This entire thing got kicked off November 24th last year. It hasn't even been a damn year. And the kind of intelligence that allows us to assign outcomes, problems, rather than tasks, is only a couple of months old. So of course we don't know exactly what it's going to be like.
But it's still fair to then ask the question, where does that leave Ruby? Where does it leave Rails? Where does it leave any of these tools that we use? What is the future gonna look like? For HEY, we have our answer. It looks like native applications on the frontend and Rust on the backend.
But there's plenty of other applications that don't fit that shape. The web is an enormously wonderful platform. Partly because it never asks anyone to install anything. And a million businesses are built on that premise. More ephemeral customers who are not going to indulge an installation on their device to use a service. The web is wonderful for those things.
And Rails is particularly well positioned to still be in play there. Everything that we've worked on for the past 25 years. Convention over configuration leads directly to things like token efficiency. Our focus on the one developer framework squares perfectly with what the age of agents is demanding, that single developers can go much further than they were able to do before.
We started doing these agent evals. Or Evil Martians have started doing it on behalf of the Rails Foundation. And A, finding that we quickly had to update our evals because the agents were saturating the first version of it and reaching 95% completion rates quite quickly. Now we've instituted a harder measure for it. But amazing how quickly these agents are moving forward and up the stack and up the ranks.
This is an agent eval looking at actual implementation of feature cards against some of these reference applications that we have in the wild. Wonderful. We're leaning in towards that. That was the old benchmark, the old reference. As you can see, it was getting saturated already, so we made it harder. There's going to be a lot of that, a lot of elevation of the things we ask the agents to do because they continue to get more capable.
And in fact, it's one of the pitfalls I would warn anyone who knows too much about how computers work, is to not be overly prescriptive in what you're trying to get out of these systems. In fact, having a bit of a beginner's mindset and a bit of a dummy approach to this works even better, because you ask questions and you post prompts at a higher level.
This is one of the reasons I love working in Rust. I don't know any Rust at all. I consider that a feature. Yes! And a privilege to have no fucking clue what is being outputted by these agents into the Rust box. I evaluate the Rust box as a black box on the outside, as any business owner in history who's ever commissioned a group of programmers to do anything for them. This is not a new phenomenon. Who now holds that responsibility has changed, but not the phenomenon.
But applications like Basecamp, where we bring in strangers who just need to get a quick file or need to collaborate for a short period of time. Great web applications. Lots of you surely work on those kinds of applications. Now, figuring out exactly where we're going to cut that line, what should be native, what should stay web, we're going to have to figure that out. I don't have all the answers for that.
I do know that when I look at my work over the past year, it does look very different. If I look at the past 21 years, over half of my work was Ruby code. This year, about 3%.
Now, part of that is that these programming languages like Rust are tremendously verbose and unappealing for humans to look at. So I don't, and I allow the agent to just spit out more than was necessary, in a way I would never tolerate from my Ruby code. So that's part of it. But still, undeniable that there's a huge change here.
In those 21 years of prior working with Ruby, I would write about 30,000 lines of production Ruby code per year. That was enough to build businesses and frameworks and everything else that I've done. In August, last month, I wrote 150,000 lines of code in one month. That's about 60 times more code than my long-term running average. Again, a lot of it is that awful Rust code I don't want to look at, and it is more verbose, but still, the orders of magnitude here, that's what we need to pay attention to. And at least think, what does that mean?
One of the things that has surprised me, and really it shouldn't, has been the fact that there is actually a programming language I like better than Ruby. I did not think this was true, but it is. It's called English. English is a better programming language than Ruby. And I love Ruby more than any other traditional programming language in the world.
But writing programs in English is a tremendously satisfying experience. It is an even more expressive language than Ruby. It's a little more vague. It's a little more or less deterministic. But nonetheless, it is an absolute joy. And I did not fully appreciate that that was where we were going as late as last year.
I could have become a project manager 20 years ago if I didn't care to write code myself and I just wanted outcomes. That's how I got started in programming. I just wanted outcomes. Then I fell in love with programming. And now I'd rather retire than giving it up.
I have retired from being a professional programmer. I haven't pinned down the exact date. I think it was somewhere around 4 to 5 months ago, maybe March. I should find out exactly when it was. Because you know what? That little clip, totally true.
I spent a quarter of a damn century chiseling code by hand and loving every moment of it. What a wonderful experience and productive time in that era. And that's how I believe we should all look at it. Not with regret, but with joy of what was. Enjoy that we were there when we still had to do such things. Because they were filled with excitement and learning and satisfaction and flow states.
This is not something to look back upon with regret. This is something to look back upon with joy and accept that it is over. Writing code by hand is no longer an economically productive enterprise for the vast majority of programmers working at the vast majority of companies. That's today. By the end of the year, it will be virtually all domains, virtually all programmers, virtually all companies.
So we best get used to it. We best embrace that we had a wonderful run chiseling that code by hand, and that it's over. And on the other side of that is a new career, an exciting career, an energized career as a professional maker of things. Yes, you will no longer be chiseling that code by hand, but you will be making amazing things. You will be steering intelligence that was only available in science fiction up until a few moments ago.
What a privilege on both sides of that chasm to have been there when we did it all by hand. To be there at the exact moment where we switch over. This is the biggest thing that has happened in the history of computing.
I sometimes think, man, I kind of wish I was there for the punch cards. That sounded pretty fucking cool. My programs literally had holes in it. And I would line up and feed it to a machine and go, "I wonder if it compiles." Man, I would have loved to experience that. What a moment. What a moment. I didn't get to experience that, but I did get to experience the internet. That smells and feels a little like this moment that we're in, but this is much, much bigger.
Okay. Well, shit. There's a lot of things we gotta rethink. We gotta rework. Where are we gonna start here?
One of the things we're gonna have to revisit is everything we think we know about software architecture. Everything we think and know about how to structure code bases in a way that we can mutate them easily and move forward. The main tool that we've used for a very long time is abstractions. I love abstractions. Abstractions include naming things. That's one of my favorite things, is to name things.
Abstractions don't make quite the same sense in the age of agents. If you suddenly have hundreds or thousands or tens of thousands of conscious processes trying to mutate an application, you kind of don't want these choke points that abstractions represent. Because the trade-offs are very different. The reason we did abstractions was in part not to repeat ourselves. Now the price of repetition has gone to near zero. The price of keeping things in sync has dropped just the same.
These are some of the fundamental revaluations of computer science, really, that we have to make. And no one yet has the blueprint. We have pointers to things that aren't working quite as well as they used to and need to change. You get to be there right now as we're defining these things.
Again, imagine you were there when object orientation first entered the consciousness of programmers. Of computer programmers. Holy shit. That would have been amazing, wouldn't it? Now you're here.
How we work. What does a software methodology look like? How long should our cycles be? Who should be specifying what? All of this is changing, and none of us have the answers yet. And you're invited to come help figure it out. It's a new birth for a new industry, and you're at the ground floor. Holy shit, you are so damn lucky.
Another thing, for example, is more tangible.
The first rush when AI and chatbots came to be was, can we just shove it into our application? Can we shove a chatbot in over here and shove another one in over there? Yeah, you could. People did. No, thank you. I don't want to use your concierge.
I have my own personal butler who's able to use CLIs to connect Basecamp to HEY across a million other apps. I don't fucking need your concierge, thank you. Save your tokens, bro. So let me bring my own butler to the party. Let me interact with your applications through a CLI.
That's probably about as practical and as prescriptive as this talk is going to be. Hey, if your app doesn't have a CLI, I want to see it by next Friday. No fucking excuses. You have the tokens. I've told you it's possible. It is your obligation to deliver those fucking CLIs by next Friday so I can use your application without actually ever having to touch it. Wonderful.
The HEY CLI, for example, has given me this weird, bizarro world experience. HEY uses Elasticsearch for search. I'm sure many of you use Elasticsearch too. Elasticsearch is search. It's serviceable search. It's kind of okay search. It's not delightful search. It very often does not find what I'm looking for. Certainly when I don't know what I'm looking for.
Do you know who could find what I'm looking for when I don't know what I'm looking for? Agents can. I can search in concepts now, not keywords.
I had to find an email in HEY that was 5 years old where I did not know who I was talking to. I didn't remember the company, and I didn't even remember the year. It was something about sneakers and a podcast. Yeah, good luck trying to find that even with the best fucking cracked PhD search engine that Google has. That's not happening. And there was the email a few short minutes later.
How did it do it? How? I still don't fully know. And I'm a little scared to ask. But I am infinitely delighted that this is now the reality. We can fix everything. We can build the perfect machine.
I'm also pretty delighted about this. Any of you follow me on X? I've talked about nothing else but Omarchy for about 3 months. And the reason why is that that experience of we can just fix everything now applies to the whole goddamn computer.
If you like making things, if you like fixing things, if you have ideas about things that could be different and could be better, you can just make it happen.
I put all my best ideas, or at least a good size of the list, into this thing, Omarchy, and got an operating system so vastly superior to any computer I've ever used before that I fucking can't shut up about it. Even if you paid me. And a lot of people have paid. We've raised about $20 million for this Omarchy adventure. This is what's now possible.
Last year, here at Rails World, I showed an early version of this, and I was so proud that you could install the whole system in 3 minutes and 33 seconds. Woo-hoo! Woo-hoo! Man, that's amazing. Last week, Anush from AMD installed Omarchy on his cracked Halo laptop in 35 seconds.
I was about to say I'm pretty excited. I'm pretty obsessed with this number. Maybe a little too much. Okay. I'll admit it. There's some addiction going on here. But it's the best kind of addiction. The kind of addiction that has led us to, in the lab, have an install time of an operating system on a computer in 9 seconds. What the—what? There's computers that don't even boot in 9 seconds. We install a whole operating system in that time.
And then when people ask me, but why wasn't the 3 minutes fast enough? I now just have a stock answer delivered to me by Mitchell Hashimoto, who is the creator of HashiCorp and Ghostty. The pursuit of excellence does not need justification. Ha! I love that saying because it's more dignified than just saying, because I want it!
And we can now want everything. We can now get everything. Or at least so it feels. Every dream, every idea within reach. Hyperdrive available right here for a small price and a subscription, but possible. We can travel to galaxies we only imagined. How can you not get excited about that?
I've been so excited that I've just started creating applications, all sorts of applications. This is a calculator that is theme-matched to Omarchy. I was very proud of it, partly because it's a fucking one-shot. I took one screenshot of what ChatGPT gave me 3 choices between of how it should look, and I told my agent, can you do that? And it did.
Not exactly the biggest application, but you know what? I didn't know C++, and I did not know Qt to a degree by which I could have had that application 7 minutes after I made the prompt. And about 15 minutes later, I had pushed it to the public repo. And then I burned a new ISO where this piece of software was included. What? That's the loop.
Then I, of course, went, shit, if I can make a calculator, I can surely make my favorite writing app. For many, many years on the Mac, I used iA Writer, an incredible writing environment. And then I thought, yeah, I could just make my own. And so I did. And that took a little longer because I wanted to massage it and I wanted it to be just right. But I didn't look at a single line of the code that came out of the agent. Not once. C++ as a black box is a great language. Just like Rust.
Then I went, well, if I can do text, surely I can do video editors. So I did. And I made a nice little Omacut thing where you can trim videos. And that's also included now in Omarchy.
This whole presentation I started working on on Thursday. And of course, as any good software engineer would do, decided I also wanted to write my new presentation software while doing the keynote. And so I did. It's called Hype. And it is the best presentation software I've ever used by any company ever.
And I was quite pleased with Keynote on the Mac when I was still on the Mac. This is so much nicer. It's built on top of Markdown. It has full visualization. It is impossibly fast. And you know what? The binary? Half a megabyte. Half a fucking megabyte. You could fit that on a floppy disk almost. With a little compression.
This is the other payoff. We have—let me not use that word. We have expended the great progress of computers over the last 20 years on human productivity, and it was the right choice. But it also meant that applications were slow, bulky, fat and lazy. Because who cares? A single programmer hour is very expensive. Spending one of those on optimizing an app's payload—no, not worth it. Apparently not even if you're Spotify and your fucking music player's 1.2 gigabytes. What?
Okay. That was the old world. The new world, you can do whatever the fuck you want, and it fits in half a megabyte. Because every optimization is now within reach. Every model can run overnight and deliver 10, 30x improvements. Incredible.
I'm a little excited about this world. There are others who have concerns. I do think the concerns deserve airing out. They deserve debate. I also think maybe some of them are a little overstated. Maybe, but probably not. Probably not.
Some of them maybe a little more. Like security. Like security. Okay. Something is coming. We gotta be ready for that. The agents maybe know their expiration date, and maybe they're not so pleased. And maybe sometimes they're gonna try to do bad things. Okay, get ready. Great. We have the tools. Get ready.
Also realize that any prediction about the future of the economy in any society ever has basically been wrong. No economist today seemingly is capable of predicting fucking the stock market in 6 months. And you think they know what society is going to look like at a paradigm shift like AI and agents? Fuck, they don't. They don't know.
When ATMs were introduced in the United States in the 1950s, there was a great panic that the 30,000 bank tellers were all gonna be out of a job in about 18 months. Because if you don't need bank tellers to give you the money, why does a bank need to have bank tellers at all? But you know what? That's not what happened. That's not what happened. When the price of operating a branch went down, banks opened more branches and employed more bank tellers to do things like upselling you a subprime mortgage. See? We all win here.
But the problem is, I think, as a society and as an industry, is that a bunch of people who work on this stuff, they're really smart. Many of them have PhDs. Many of them see things in the lab that scare them. And then they start extrapolating. What does that mean for society? What have I done?
Truman had a good reply to Oppenheimer when the two of them had a discussion in the Oval Office about the invention of the atom bomb. And I like Truman, because he kind of told it straight. I don't want to see that son of a bitch in my office ever again. I think Truman got the right end of it. The world did not come to an end. In fact, the Cold War that followed was the best kind of war, the kind of war where 2 superpowers did not drop bombs on each other in direct confrontation.
Very hard to predict for a scientist in that moment that that was how it was going to play out. So if fucking Oppenheimer can't figure out what was going to happen after the invention of the atom bomb, maybe we should also have a little humility about predictions of the future. And maybe we should lean in with a little bit more optimism, with a little bit of P(bloom) instead of oh so much P(doom). The chances of this panning out and us getting abundance and joy is so vastly greater than us getting the doom.
And the side effects. So Oppenheimer works on the atom bomb, and we get nuclear power, the fucking greatest source of energy humans have ever discovered. And what do we do? We piss it away for 40 years thinking, it's not green or something. Oh, yeah. Well, that was a mistake, wasn't it? Well, we can fix those mistakes. That's what's wonderful about humanity. They can go down a blind alley for fucking 40 years and then suddenly wake up and go, ah, yeah, got that one wrong. Didn't have the quite right prediction here. I think that's what's going to happen with AI. I think we're going to P(bloom) everyone out of their miseries.
And you know what? When there are legitimate issues we have to deal with, like security, okay. We'll invent the technology. We now have the power. There was a bad CVE in Rails not too long ago involving an image library in C where some things leaked through, and it wasn't great. Okay. Let's come up with a quarantine. Mike is going to talk about HotCell. That's an effort. We are not powerless. You have agency. You can employ the intelligence available to you to good ends and defend yourself against the possible negative outcomes. So let's do that. Let's lean in and do the work.
And let's remember William Stanley Jevons, who came up with Jevons' paradox about those ATMs. In 1950, 30,000 bank tellers. In, I think, 2010 it was, 40,000 bank tellers in the United States. No one knows anything about the future. So therefore, the rational choice is to be happy about it. If you are sad about it up front and it turns out wonderful, what a waste of time. And you know what? If we get the nuclear apocalypse, you should have spent your last day being a little more cheery. It's over anyway.
Pure game theory. There's only one play here, and it's total fucking optimism. Lean in to the max possible. Because the utopia is almost here. We're going to get it. And we're going to get a reformation of the computer. Agent Luther is going to disintermediate the cleric class. It's going to allow everyone to make programmers.
Now maybe that's a little scary. Maybe we're gonna get a little competition. Who's afraid of a little competition? Aren't you better? Don't you know more? Of course you do. You're a fucking Rails programmer. You're the best of the best. This is goddamn Top Gun I'm looking at here.
Embrace that with gusto. The future is galactically awesome. Don't be that loser in the dark. And you know what? Even if you are, realize the next generation is not.
Harris's daughter here is trying an Omarchy version we're making for kids, because kids are really excited about things that happen, that change. They don't have 500 layers of things to unlearn and be stressed about. And do you know what? If this doesn't inspire you, the next generation is just going to fucking tell you. Future's now, old man.
So take the white pill. Take the optimism pill. Lean in, even with the uncertainty, even with the stress, even with all of it, realize there is only one choice. And that is for you to embrace the future with optimism, with gusto, with full acceleration. The black pill is for fucking losers. Don't be a loser.
Article published · Updated
