Hacker Remix

Show HN: A big tech dev experience for an open source CMS

52 points by randall 2 days ago | 28 comments

Hey HN! We're building an open-source CMS designed to help creators with every part of the content production pipeline.

We're showing our tiny first step: A tool designed to take in a Twitter username and produce an "identity card" based on it. We expect to use an approach similar to [Constitutional AI] with an explicit focus on repeatability, testability, and verification of an "identity card." We think this approach could be used to create finetuning examples for training changes, or serve as inference time insight for LLMs, or most likely a combination of the two.

The tooling we're showing today is extremely simplistic (and the AI is frankly bad) but this is intentional. We're more focused on showing the dev experience and community aspects. We'd like to make it easier to contribute to this project than edit Wikipedia. Communities are frustrated with things like Wordpress, Apache, and other open source foundations focusing on things other than software. We have a lot of community ideas (governance via vote by jury is perhaps the most interesting).

We're a team of 5, and we've bounced around a few companies with each other. We're all professional creators (video + music) and we're creating tooling for ourselves first.

Previously, we did a startup called Vidpresso (YC W14) that was acquired by Facebook in 2018. We all worked at Facebook for 5 years on creator tooling, and have since left to start this thing.

After leaving FB, it was painful for us to leave the warm embrace of the Facebook infra team where we had amazing tooling. Since then, we've pivoted a bunch of times trying to figure out our "real" product. While we think we've finally nailed it, the developer experience we built is one we think others could benefit from.

Our tooling is designed so any developer can easily jump in and start contributing. It's an AI-first dev environment designed with a few key principles in mind:

1. You should be able to discover any command you need to run without looking at docs. 2. To make a change, as much context as possible should be provided as close to the code as possible. 3. AIs are "people too", in the sense that they benefit from focused context, and not being distracted by having to search deeply through multiple files or documentation to make changes.

We have a few non-traditional elements to our stack which we think are worth exploring. [Isograph] helps us simplify our component usage with GraphQL. [Replit] lets people use AI coding without needing to set up any additional tooling. We've learned how to treat it like a junior developer, and think it will be the best platform for AI-first open source projects going forward. [Sapling] (and Git together) for version control. It might sound counter intuitive, but we use Git to manage agent interactionsand we use Sapling to manage "purposeful" commits.

My last [Show HN post in 2013] ended up helping me find my Vidpresso cofounder so I have high hopes for this one. I'm excited to meet anyone, developers, creators, or nice people in general, and start to work with them to make this project work. I have good references of being a nice guy, and aim to keep that going with this project.

The best way to work with us is [remix our Replit app] and [join our Discord].

Thanks for reading and checking us out! It's super early, but we're excited to work with you!

[Constitutional AI]: https://www.anthropic.com/research/constitutional-ai-harmles...

[Isograph]: https://isograph.dev

[Replit]: https://replit.com

[Sapling]: https://sapling-scm.com

[Show HN post in 2013]: https://news.ycombinator.com/item?id=6993981

[remix our Replit app]: https://replit.com/t/bolt-foundry/repls/Content-Foundry/view...

[join our Discord]: https://discord.gg/TjQZfWjSQ7

jmuguy 2 days ago

Maybe its just me but I can't figure out what exactly this is. I tried your demo, gave it a Twitter user. And its now generating blog posts in the "voice" of that Twitter user? And this is a tiny first step to... what? I know this sounds antagonistic but I'm just genuinely confused.

randall 2 days ago

Totally fine, thanks for trying it out.

The idea is this "identity card", which is super early in our dev process. The goal is to make it so you can have your voice as a thoroughly vetted example, and then be able to use that to finetune LLMs, or use it at inference time to enhance LLM outputs to sound more like you.

We'll be building out a full CMS, and shipping parts of it over time.

Thanks again, it's super early but we're working on it. :)

jmuguy 2 days ago

Ah, I'm not much of an AI person so I'm probably just lacking context. Regardless I do think it would be helpful to point out - your second paragraph refers to "the tooling" and that you're focused on "showing the dev experience and community aspects". I don't really see how your site or the demo you're showing actually highlights either of those things.

randall 2 days ago

Yeah, honestly this might have been too early of a launch, but kind of just want to put something out and then iterate.

fallinditch 2 days ago

> you can have your voice as a thoroughly vetted example, and then be able to use that to finetune LLMs

You have piqued my interest.

Can you explain in more detail how a user will fine tune a model. Will the LLM just replicate a user's conversational style? Or will the user be able to set deeper parameters, for example to define goal-oriented outputs for agentic actions?

randall 2 days ago

So the goal initially is going to be to create a "constitution." Basically a document that outlines your voice, or whatever.

From there, you can create examples for every sentence and classify them... ie "this is a good example" or "this is almost a good example."

Then, after you have a good enough enough number of human versions (idk this number yet), you can use those to create dozens more examples, refine, etc... then at some point you've thoroughly described a document enough that there's no more delta. Then you can use the document to train / finetune / infer.

The effects are that the "instincts" of the model (if on the training / weights) or the "thoughts" of the model (at inference time) are closer to a well vetted baseline.

rbalicki 2 days ago

I'm incredibly excited for this launch! Bolt Foundry is the first company to use [Isograph], which is a framework for building data-driven apps that are performant and stable out of the box. It does so by making heavy use of a compiler and of generated files, and essentially being an incremental computation engine for your UI. Your UI becomes performant by doing the least amount of unnecessary work.

And Bolt Foundry is truly the perfect first adopter! As "A big tech dev experience" implies, they're very focused on providing great DevEx. And a bunch of them previously worked at Meta, where they used Relay, and understood the powers that it provides.

Isograph aims to provide an even better developer experience (see the YouTube link), and give you the things that previously were only feasible at FAANG-co scale: e.g. even in a very dynamic app, fetching just the JavaScript and data you end up using.

So far, they've had a great experience using Isograph!! You should also [give it a try] or [contribute].

[Isograph]: https://www.youtube.com/watch?v=sf8ac2NtwPY

[give it a try]: https://isograph.dev/docs/quickstart

[contribute]: https://github.com/isographlabs/isograph

graypegg 2 days ago

Interesting! I think the idea of a "tone" in your content being enforced by a sort of... prose linter makes a lot of sense.

You can assume 1 person with at least some practice in writing will have a clear tone and presence in the things they write, but this would be rather useful for corporate communications, especially when simpleton engineers (such as myself) try to write blog posts for a company engineering blog. If there was something nudging me towards the style of the blog, without steam rolling me constantly by rewriting everything, I would like that.

For the CMS, (I think you're building a CMS?) I think building around THAT use case would be an awesome selling point. Let content folk spawn special ephemeral links that let someone write up a draft without having to set up an account. Have some sort of approval process in there too, that could be configured on that "magic link", so I don't impulsively publish to everyone.

Make it super easy to just stumble into writing something the company can actually communicate with, right away. There's a lot of passionate people at companies, with 0 writing experience that would lend a lot of authenticity to blogs/engineering comms/recruiting ads, if they they had the bumpers in their bowling lane.

randall 2 days ago

precisely.

that’s exactly where we’re headed.

thanks!!

Etheryte 2 days ago

For your consideration, the main page you've linked to is nearly unusable on mobile, might want to look into that.

randall 2 days ago

We will fix right now.

graypegg 2 days ago

The tweaked typography is looking better!

You'll need to toss on

    <meta name="viewport" content="width=device-width, initial-scale=1.0" />

and add a small amount of padding around the body to make the text scaling pixel density agnostic though.

I would also remove the overflow:hidden on body, and remove the fixed height on #root. Nothing needs to be a scrollable container here, the viewport can just scroll over the body. That's causing the bottom ~2 paragraphs to be "stuck" under the bottom edge of my iPhone 13 mini in safari. Ideally use `min-height:100dvh`.

randall 2 days ago

Just shipped it. Will get the padding going too.

fragmede 2 days ago

https://imgur.com/a/n1zVO1n

I can't read that and pinch to zoom has been disabled.