You’re tired of building the same components and CSS for every client, so you cook up a system. It’s pretty slick, and you make your money from clients anyways, so you opensource the design kit.
At first, just a few stars roll in on the Github repo. But then, thousands, even more. And you’re now the top frontend CSS repo.
Everyone’s using it. So, you build a premium components library on it. Everything you’d need to build a design website, a web dashboard, even a progressive web mobile app. And you throw up a Stripe link and see if people will pay.
A few do. And then a few more. Again and again your phone buzzes with the notifications.
Soon, enough are coming in you (and your two buddies you built it with) quit your jobs. You eventually hire 4 more engineers to help build the next version.
Life’s good. You made it!
Something useful, something people are willing to pay for, and a base layer common good released opensource to the public that all can use for free.
A perfect equilibrium. All is right with the world.
Until one day, a few years later, the growth curve started tilting the down, and never pulled up.
This is the story of Tailwind CSS, and what it’s near total destruction from the AI revolution can warn us about the state of opensource software going forward.
Hire me as CTO or tech lead for your next project.
Focus on growing the business, never worry about the tech again.
Slide into my DMs on X or Substack. Select portfolio at Deca Labs.
Down 80%, with no sign of stopping
All the talk last week was of Tailwind CSS, the much beloved utility styling frontend library which for years has grown in popularity to become one, if not the, dominant component frameworks for web development.
And yet despite their generational run, traffic and revenue were down 80% since the launch of AI. They just had to lay off 75% of their engineers.
The second order effects of making software effortless to write with AI means the end of an entire class of opensource software (OSS) projects.
Without devs organically going to the documentation site, there was no way to upsell their premium components packages. And with AI being able to generate thousand line complex components with for $0.12 of credits anyways, why pay $65 for components up front?
So far, the AI generated components have lacked the polish of Tailwind’s premium offerings, but how long will that edge last?
Use Poaster’s custom AI model to pull the best quotes from your long form writing, and automatically post them to your socials.
Skip the army of virtual assistants, start today for $1/month.
Even in recent months, some corporate executives have gotten the whiff of AI eating the world, and decided that continuing to support existing libraries and maintainers isn’t a priority anymore.
I’ve seen industry wide titans and 15 year veterans of US big tech software companies have their opensource projects cancelled, and managed out without severance, despite building and supporting libraries used in almost every Android app used on Planet Earth.
Within just a few short years, the economics of opensource software has gotten turned on its head.
Thank you to the readers who pay to make this newsletter possible.
Ready to turn your life around? Subscribe & start here and report back your wins.
– Fullstack
This Substack is reader-supported. To receive new posts and support my work, consider becoming a free or paid subscriber.
Winner takes all, even more
Opensource libraries are abstractions.
They hide some of the complexity of doing something, and make it easier for devs.
Instead of writing every byte yourself, an http client can take a request in JSON and then handle the complexity of TCP for you.
Existing top libraries per category have dominant downloads and usage, sometimes 10x or 100x the next similar library that can do the same thing in that language. Devs flock to what works, is proven, and has enough momentum and backing to remain well supported.
In previous years, a new library would get published on Github, and some of them would organically accelerate in popularity until they were the default library among developers.
From word of mouth, Hacker News, Product Hunt, Reddit, Slack, and a variety of mediums, an upstart could take on the most established library and have a possible shot at winning.
Devs would get loured in by the hype, skim the README and docs to get a feel for it, and maybe try using it next project.
With AI, this organic discovery doesn’t happen anymore.
AI spits out working code, you merge the PR and move on.
Devs often don’t even check the new libraries AI is pulling in to their project, because if it works? Ship it!
And AI, unlike the fickle programmers of old, is not easily swayed by the newest shiney library to rack up the stars on Github. AI predicts the most likely successful code, which means the status quo.
While launching a new opensource project was as difficult as launching a SaaS before, now it is much, much harder. There’s no SEO hacker technique to insert your new library into the AI foundational model code generation module.
The status quo has been crystallized, this may be the beginning of the Long Night, to borrow a concept from John Robb.
Are you looking for 1:1 coaching (tech career, wifi-money, Canadian real estate)?
DM me on Substack or X with your topic, I’ll see if I can help. $120 for 30 mins.
– Fullstack
Brief relief for Tailwind CSS, but the bleak trajectory holds
Given the widespread adoption of Tailwind CSS industry wide, there was an outpouring of support for the beleaguered founder, Adam Wathan, who admitted the depressing nose dive the company was in across his podcast on X and while closing Github Issues from community members who wanted to make it even easier to never visit the Tailwind CSS documentation site (& be upsold on their premium components).
(and as usual, OSS consooomers showed up on their best behavior)
If you have the time, I’d recommend giving the podcast a listen. It articulates well the black pill perspective of the impact of AI on coming millions of people.
It won’t be just Uber drivers getting replaced, but potentially millions of white collar previously prestigious jobs.
Thankfully, Tailwind appreciators across the industry forced their employers to do the right thing and step up as sponsors to bolster the project and keep it from fully collapsing. MRR jumped from $75k to $130k in a week, mostly from new corporate sponsors.
But despite a brief jump in sponsors for Tailwind, the bleak trajectory holds.
Opensource software has suddenly transitioned from simply being very difficult to monetize, to now nearly impossible.
Old monetization models of either 1. premium features (Tailwind) or 2. premium support (Jetty), are quickly being phased out. AI can generate any premium features on demand for $0.12 in credits, or answer any premium support question or problem in seconds.
This will leave only hobbyists or corporate opensource maintainers, no more self-employed solo or tiny software shops.
Pour a cold one out. It truly is a sad day, and yet another unexpected casualty of the AI revolution.
But like any revolution, it’s best to not get caught on the wrong side of it when the bullets start firing.
The time to learn how to be ruthlessly effective with AI was yesterday.
Time is ticking, and either we’ll all be out of a job... or only those that can use the AI to have 10x impact will remain employed.
Given what I’ve seen, I’m still betting on the latter. Time will tell.
If you want to survive, all I can say is this:
Continue reading this post for free, courtesy of BowTied Fullstack.
