Beaker Browser MVP Review: 0 use cases + 4 years of engineering = 0 users

Beaker Browser is an open-source “peer-to-peer browser for Web hackers” project led by Paul Frazee and Tara Vancil. Here’s the kind of description you can find on the project home page:

The Web should be a creative tool for everyone. Beaker brings peer-to-peer publishing to the Web, turning the browser into a supercharged tool for building websites, files, apps, and more.

Ok… what does any of that mean in terms of an actual use case? We’re off to a vague start.

Here are various features listed on the home page:

• Deploy instantly
• Co-host sites peer-to-peer
• Build peer-to-peer apps with new Web APIs
• Explore files using a fully-featured filesystem
• An integrated terminal
• An integrated editor for web page source code

Hmm, it’s definitely smelling like it’s just an engineer building stuff that engineers like to build, regardless of whether there’s a plausible use case.

How much work went into this?

According to Github, it was over 4 person-years of coding work, beginning in May 2016 until being abandoned in July 2020.

Last year, while the project was still in active development, I had this back & forth with Paul on Twitter:

Paul: If the Web includes an open serverless platform, the creation of networked apps will be accessible to all users. Forums, collaborative spreadsheets, chat rooms, and etc will be as convenient to create as documents in a word processor. And added benefit is radical interoperability: by sharing the user’s personal data store, these applications will be wired into a shared network

Liron: What’s a specific use case for which today’s best forum, spreadsheet and/or chatroom solutions are inferior to what you’re imagining? What’s a specific use case that would be a lot better with increased interoperability?

Paul: I’ll answer both your questions here. All software has room for innovation. In the current world, users are disempowered to do it. We get Twitter as it’s given to us. We get Reddit as it’s given to us.

Liron: What’s one plausible example of a use case Twitter or Reddit is inadequate for, and the corresponding improved use case on your envisioned platform?

Paul: At some point I have to ask who here lacks imagination! You’ve never wanted a better DMs system in Twitter? More social moderation tools? A reply system with threads? Better ways to track reputation? When people think about self hosting a forum so they can get tools that reddit doesn’t give them, they’re our audience. Custom deployments require server admin. On the p2p web, the idea goes, they won’t

Liron: What’s an example of a specific reason someone wants to set up a forum but isn’t happy with neither Reddit nor any of today’s other forum software?

[No reply]

Liron: I just watched Tara’s JSConf 2018 talk. To build “better moderation tools for Twitter” (mentioned at 24:05), wouldn’t hardest parts be (1) chicken & egg problem of getting users and (2) scaling backend to millions of users? So then Beaker Browser isn’t the needed tool here.

[No reply]

What’s the killer app?

In the above back & forth, I was basically asking Paul to tell me what’s a plausible first “killer app” of his platform.

Paul’s tweets are instructive because they’re the telltale sign of someone who thinks they have a strong value prop, but actually just has a big list of weak value props, which is useless. No matter how many weak value props you list, they don’t add up to a strong value prop; they’re only as strong as the strongest one in the list. His list is:

• A better DMs system in Twitter
• More social moderation tools
• A reply system with threads
• Better ways to track reputation
• Deploy without server administration work

If we pick one value prop from the list, e.g. “a better DMs system in Twitter”… well, what are we even talking about? How is it a better DMs system? I guess because it’s peer to peer and people control their data… but now we’re talking in abstractions. We’ve gotten too far removed from a specific description of how a specific person’s life is getting better before vs. after.

Right now, if I didn’t want Twitter to see my messages, then I’d just use a platform that end-to-end encrypts messages. And maybe one day Twitter itself will add an option to end-to-end encrypt DMs. It’s so weak to just answer my question by saying “a better DMs system in Twitter”. It only looks like it might be an awesome answer because it’s followed up by a bunch of other things in a list, that together make you think about a big abstract vision of a “better web”, or something. This kind of abstraction trap is why I insist that a Value Prop Story must be a highly specific story, not merely an abstract story.

Paul was interviewed on the Lunar Society Podcast a few months ago (June 2020) and the discussion turned to what’s a killer app for Beaker Browser. This excerpt from 16:45 is revealing:

Interviewer: Are you guys thinking about applications that could be built on this protocol, using this new paradigm, that just can’t exist in the existing ways of thinking?

Paul: Yeah, you’re also touching on what a lot of people talk about, which is the “killer app” question. In my opinion, especially in the early lifecycle of this, the killer app is the hackability. There are a lot of developers out there who like being able to do this sort of thing. The ease with which we can both take existing code and change it, or just create a new application and put it out into the world, has never been matched in my opinion whenever you use a technology like this.

And this connects to the fact that when you do it this way, you also get rid of what you call “dev ops” in the tech industry, which is the process of setting up and maintaining and then paying for a server. Well a peer-to-peer system is designed to more or less handle scale automatically. As more people get involved, they share bandwidth. So as someone is publishing the software, you’re able to put the software out there without having to think about it anymore.

If the differentiator is the ease of putting a new application out into the world, that value prop is competing against the value prop of e.g. today’s full-stack web frameworks, right? If that were true, then there should be some kind of demo or proof of that, like Beaker Browser competing with other frameworks in ease of building TodoMVC. Or is it competing against the value of easy-to-use hosting platforms like Firebase? Again, no head-to-head demo proving superiority.

It seems they’d rather not stoop to naming and validating a specific killer app. They’d rather keep thinking of Beaker Browser as a big abstract ball of irreducible awesomeness.

Yet another developer platform bloated MVP

This Bloated MVP is structurally very similar to another one I’ve written about: Axiom. Like the Axiom team, the Beaker team doesn’t have a clue what Beaker’s killer app could plausibly be, but they think they’ve found an escape into a level of indirection by saying “developers will love it and build a killer app”. But what killer app might the developers plausibly build?

Stop hiding behind your level of indirection and think about the whole value chain. If you don’t think all the way down to a single plausible example of an end-to-end value prop story, then you’re overwhelmingly likely to end up with zero users. You didn’t need four years of writing code to learn that you have no plausible mechanism of creating value. It was already obvious on Day 1 that the Value Prop Story sanity check was not passing.

Are these Bloated MVP reviews getting a bit repetitive? Are you wondering why is this guy just sitting at his computer in his underwear, not working, not spending time with his wife and kids, just dunking on people from Twitter?

Look, I meet new founders every week who are working on their own Bloated MVPs, super high on their own supply, and I’m having trouble getting them down to earth and realizing they don’t even have a freaking Value Prop Story. So maybe when they see that there are dozens of posts on this blog, dozens of lemmings in front of them in line who have already marched off the cliff, they’ll think twice and save their precious supply of IQ-hours for a project that actually creates value in the world.

I’ll leave you with one more back & forth, wherein my reply never received a followup reply.

Update Jan 2021: Paul graciously replied on Hacker News:

I actually think this [Bloated MVP critique] is fair and I’ve been talking pretty openly lately that we failed to identify a solid market for Beaker. This is why we’ve turned our focus lately to other tools in the ecosystem and been looking more closely for a PMF [product-market fit] for the p2p tech. Happy to share more on HN but I’ve talked a lot about it on my YouTube channel already