jessmartin

Start with Jayber Crow.

Then read Hannah Coulter.

Then the whole Port Williams collective will open before you like a beatific field.

Philosophy and CS double major here, love reading old books. Read a lot of theology and ancient literature. Homeschooling my kids, part of the curriculum is learning Latin.

Waiting for “JSONL is all you need for durable workflows.”

Agree. I ~never see that stuff.

I love the idea of using a shared event log for coordination. Smart!

I have a Symphony-style[1] factory, which keeps all the context in a single session, but I want to start splitting into stations with separate sessions, and I hadn’t worked out how to do communication between sessions.

[1] https://github.com/openai/symphony

Really cool to see JSON Canvas getting attention. (Hi, kepano!) It was one of the inspirations for OCIF (Open Canvas Interchange Format), which we've been working on since JSON Canvas came out.

JSON Canvas nails the simplicity-it's easy to read and easy to implement. We wanted to build on that spirit while tackling some of the challenges showing up in this thread: nested canvases, extensibility for custom app data, text styling, coordinate systems, and round-tripping between different canvas apps without data loss.

OCIF v0.7.0 just came out. It's designed to be an interchange format — so different canvas tools (Excalidraw, TLDraw, Obsidian, etc.) can export/import each other's canvases.

Some highlights:

- Extensible: apps can attach their own data via extensions, so nothing gets lost even if the features aren't supported

- Nested canvases via parent-child node relationships

- Local coordinate systems (addresses the pixel positioning concerns raised in another comment here)

- Text styling, viewport control, and more via built-in extensions

If JSON Canvas isn't quite meeting your needs, OCIF[1] might be worth a look.

https://canvasprotocol.org

strong agree. I always have the LLM put an actual markdown doc in a docs/plans/ folder before starting work. I often, but not always review it.

Aside: it also helps for code review! Review bots can point out the diff between plan and implementation.

Some examples for the curious: https://github.com/sociotechnica-org/symphony-ts/tree/main/d...

none yet!

Not a small amount :)

I spend $140/mo on Anthropic + OpenAI subs and I use all my tokens all the time.

I've started spending about $100/week on API credits, but I'd like to increase that.

I got my own level 8 factory working in the last few days and it’s been exhilarating. Mine is based on OpenAI’s Symphony[1], ported to TypeScript.

Would be happy to swap war stories.

<myhnusername>@gmail.com

I wonder if this is just a byproduct of factories being very early and very inefficient. Yegge and Huntley both acknowledge that their experiments in autonomous factories are extremely expensive and wasteful!

I would expect cost to come down over time, using approaches pioneered in the field of manufacturing.

They have a Products page where they list a database and an identity system in addition to attractors: https://factory.strongdm.ai/products

For those of us working on building factories, this is pretty obvious because once you immediately need shared context across agents / sessions and an improved ID + permissions system to keep track of who is doing what.

I’ve been building using a similar approach[1] and my intuition is that humans will be needed at some points in the factory line for specific tasks that require expertise/taste/quality. Have you found that the be the case? Where do you find that humans should be involved in the process of maximal leverage?

To name one probable area of involvement: how do you specify what needs to be built?

[1] https://sociotechnica.org/notebook/software-factory/

Wow. This is exactly my kitchen faucet. Which makes sense, because my house was built in 1961.

Works great. Amazing how durable the faucet is!

This! It’s both-and. Literacy has been undeniably good, but we rarely consider the consequences of widespread literacy.

There’s a way of knowing something that can be recalled orally from memory that is different and valuable. But we even measure it using a yardstick for written knowledge (accuracy, breadth, etc).

I believe this overemphasis on written knowledge (really, it’s implicitly a denial that any other type exists) is part of what drives the hysteria about LLMs ending the world. LLM doomerism has to believe that written knowledge is at least the most important if not the only necessary form of knowledge.

Can you say more? Link?

Same. Was very excited about MCP but Claude code + CLI tools is so much nicer.

Has anyone put together a comparison table for Desktop apps?

I’ve started using conductor.build and it feels nice, but would happily evaluate others.

Relay user here! It’s great. Quite reliable for an early product.

One of the authors of the Landscape here. Glad you found it helpful!

Could be the prompt and/or tool descriptions in whatever tool you are using Claude in that degraded. Have definitely noticed variance across Cursor, Claude Code, etc even with the exact same models.

Prompts + tools matter.

curious: did you literally use the Local-first Landscape[1] to do the evaluation? (full disclosure: co-author of landscape)

[1] https://localfirst.fm/landscape

Just saw the talk on it at Local-first Conf. So many snyc engines these days. LiveStore is exploring an interesting space in the idea maze: marrying event sourcing with sync engines.

I couldn't believe how robust LiveStore is already. I've been trying it out for a new project over the last few weeks and it's so smooth.

This seems like a good flow! I end up adding a "spec" and "todo" file for each feature[1]. This allows me to flesh out some of the architectural/technical decisions in advance and keep the LLM on the rails when the context gets very long.

[1] https://notes.jessmart.in/My+Writings/Pair+Programming+with+...

It's so good!

Not only can you get reverse chronological feed, you can build your own feed with custom logic with a few clicks and share it with other people!

https://docs.bsky.app/blog/feature-skyfeed#what-is-skyfeed

Highly recommend the starter packs for finding people!

I post about local-first, future of computing, protocols, education, and, recently, AI engineering.

Once they add gifs, I’ll post more “build out loud” threads.

https://bsky.app/profile/jessmart.in

Over the past few years, I've worked on and used a bunch of different frameworks to try to solve for distributed state. Jazz feels like it has the right "shape" - the right primitives to quickly build an app with magical powers like offline, real-time collab, etc.

If you can push the distributed state problem below the framework level, it becomes dead simple to build apps, which feels great.

Love the name.