Why the Universe will launch as a public beta

For many years now, I've been a student of software. I love what good software can do, and what a good service can do for the people who rely on it. I've also learned what happens when a product people depend on is rushed to market: all sorts of things go wrong, and the people relying on it are the ones who feel it.

I can't be a good steward of your business without telling you that.

So here's where things stand. Some of my clients seriously needed the Universe before it was ready, and most of my coding and agency clients through the Forge have been working on a private beta: a restricted version, shaped around their projects. After launch, they'll move into the full platform over time. Everyone else will come in when the Universe launches, with Intrepid and the EGT* Studio inside it, and it will launch as a public beta. It may stay that way for a long while, because I can't yet fully know how you'll use it.

This article is about why I'm choosing that, why the platform is built the way it is, and what your reservation will hold.

A public beta, in groups

A beta is the stage where real people use something before it's finished. A private beta is invite-only and restricted. A public beta is open to anyone who wants in.

Ours will work like this. Anyone can reserve a place. When the gates open, I won't be opening the floodgates. People will come in groups, mostly in the order they reserved, and each group will get settled in and watched before the next one arrives, so what the first group teaches us can help the second. The Studio exists, but it hasn't launched, and that's exactly what reservations are for. A reservation is your place in line for launch, not a login for today.

Google did something like this with Gmail in 2004. It started with about a thousand people and let more in round by round, because the machines behind it couldn't yet carry millions. One history of Gmail calls that slow start "born of necessity." Mine is, too.

Nothing talks to each other

The biggest problem I see in businesses, small, medium or large, is that nothing seems to talk to anything else. Even the tools sold to connect everything tend to leave things disjointed. I've watched teams spend hundreds of hours stitching an old system into a new one, or just moving their data across. And many of my clients keep older systems running on purpose, because they're too expensive to replace, or because nothing new truly matches what they do or how fast they do it.

That taught me something. Some of the most dependable software ever written comes from an older, procedural tradition: the world of COBOL and IBM's RPG, often with no graphical interface at all. Those programs excel because each part does one job, inside clear boundaries, and does it well. Much of the modern web was built on a different dream, where anything can be made to do almost anything. That freedom has produced brilliant work, but without boundaries, one small change can ripple out and quietly stop other features from working, and two pieces of code end up doing the same job. Agentic coding, and vibe coding, can make that worse, faster.

When I built the Universe, I carried those older lessons into something new for the web: components that each do a single job well, inside their own boundaries, so the whole can grow while one change is far less likely to break the rest. That's the philosophy. The details can wait for another day.

What I won't promise

AI is now part of how businesses get built. It's also part of how attackers work, and agents and attackers are moving fast. A lot of people launch by saying, here's my software, it'll work perfectly, you'll love it, nothing will go wrong. I'm not going to say that.

Look at PHP, the language behind most of the web. In a single week this October, two critical security holes came to light in it, the kind that can let an attacker take over a server, and they were found with the help of AI. The fix shipped within a day. If something that much of the internet stands on can be surprised that quickly, I'd be foolish to tell you mine can't be.

I'm also one founder. Even with great agents and a team around me, I don't have hundreds of security professionals standing behind me. I want this to be perfect, and I know that wanting it isn't the same as knowing how it will hold up once real people arrive.

I created something I believe wholeheartedly in.

The lab and the wild

The Universe is built on history, on what businesses actually use, on research and testing, and on what I've learned alongside the people I've partnered with. It's built well. But building something in a lab is a very different thing from releasing it to the public. People will use it in ways I never imagined, and some of those ways will teach me things no test could. I'm far from the first to learn that.

Even Twitter. For a few years, one of the most famous error messages on the internet was a cartoon whale. When Twitter had more people on it than it could carry, the whale showed up where your timeline should have been. In the early years it showed up a lot, and it came back during the 2010 World Cup. Then Twitter's engineers owned their mistakes, stepped back and rebuilt what sat underneath. Three years later, a record 143,199 tweets arrived in a single second. Twitter wrote that its users "didn't experience a blip." Even Twitter, with a whole engineering team, could only learn its real limits from real people using it.

Even Apple. In 2012, Apple replaced Google's maps with its own in iOS 6 and sent the new Maps to more than 100 million devices. Developers later told CNET they had flagged problems during the iOS 6 betas. Tim Cook wrote a letter to customers that said, plainly, "we fell short." When Apple rebuilt Maps years later, it said the new version would roll out "section by section." Testing only helps if what testers find can change the plan, and even Apple came back to the same answer: listen to the people using it, and grow in steps.

Why reserve a beta at all

You might be wondering why you'd want a place in a beta. A beta doesn't mean the platform is unstable, or built wrong. It's built well, in a lab. Once it's in the wild, I'll be making rapid changes and growing it, and I don't want that to hold you back from stepping into a whole new way of running your business.

I've seen what happens with my own clients. Even on a day when something is temperamental, or the system has to come down for maintenance, using it changes how they feel about running their business. When I release their software, they're astonished at how quickly things can be changed and updated. They go from hating whatever they had before, and I've been told this many times, to loving their new website, their new app, their new system. That tells me I'm doing something right.

When your group comes in, you won't be waiting on the sidelines for a finished product. You'll be building it with me. What you run into, what confuses you, what you wish it did: that will shape what gets fixed first. Founding members hear from the team directly, and what you tell us goes into the work.

Being early will mean a few honest things. Features will change, and some won't work the way you expect. Some days, something will glitch in a way nobody has met before. Intrepid and the Studio are built on AI, and AI can surprise you, sometimes brilliantly and sometimes by getting it wrong, so check what it makes before you rely on it. The Studio is built so nothing is generated, spent or shipped without your yes. Keep your own copies of anything that matters to you. I can't guarantee it will all work perfectly, or work the same way from one month to the next, and using it will always be your choice. The formal terms live in the EGT Codex. This is the plain-spoken version.

What a reservation holds

A reservation works a lot like a pre-order. It holds your place, and it tells me how many people to get ready for, so I can size each group before it comes in. When Valve took reservations for the Steam Deck, it said the point was to "ensure an orderly and fair ordering process." That's the spirit here: a line that's fair and easy to understand.

Here's the plain part, said once. Your reservation is non-refundable, credited in full toward your launch. Think of it as part of your first month, paid early. You're welcome to cancel your reservation at any time, but it isn't refunded. I'm a small operator, and I'd rather tell you that clearly now than surprise you later.

Your place never expires. You paid for it, and it's yours. When your group is ready, we'll email your invitation. Invitation links expire for security, and getting a fresh one is easy. If the timing isn't right for you, you can wait for a later group.

Not ready to reserve? That's fine, too. Join the newsletter and get the email when self-serve opens and the platform is ready for general use.

Love your business again

Here's what I'm really after. Too many people go to sleep with a migraine because running their business means fighting their technology every single day, forcing square pegs through round holes and calling it a workflow. I want that to end. I want technology to work with you and for you.

It doesn't have to be perfect. I just want you to love your business again.

The partner I lean on in my own life is the Holy Spirit. I can't be that for your business, and I'd never pretend to. What I can give you is a partner for the technology side of it: your website, your data, your workflows, the digital work of every day, made easier, calmer and a lot less infuriating.

I don't know everything that will happen once the doors open. Nobody does. But launching in public beta lets me bring this to you far sooner than I otherwise could, and let you grow with it. EGT was founded on hope for a brighter future. This is me sharing that hope with you, so you can love your work again, chase the business you've been dreaming of, and make the impossible possible.