Jesus Moreno · JMCodes

The Tamagotchi is really just a loading spinner.

OpenCode SDK is basically an agent framework, think VoltAgent or similar.

You plug into OpenCode’s existing agent loop and build your own UI on top.

On my 'desktop', I used the same loop to build an AI-powered graph editor.

Updated Tamagotchi screenshot showing the refined edge-to-edge interface and animated pixel character reacting to activity state.

Quick update: I tightened the Tamagotchi UI and wired mood + animation to live OpenCode activity like reading, patching, replying, and retry. New screenshot below. Details coming soon. #OpenCode #BuildInPublic

Screenshot of a compact Tamagotchi app in a desktop AI workspace: pixel pet in a 12x12 SVG grid, mood badge set to SAD, three vitals (F/E/J), and Feed/Play/Nap controls.

I recently switched to OpenCode as a long-time Claude Code user, and the biggest unlock is this: you can embed AI into any application layer, from silly experiments to real production UI. Built this reactive Tamagotchi as a tiny proof. Full build thread soon.

I help founders move from prototype to production without losing momentum. Engagements focus on the critical path: architecture, implementation quality, and workflows that hold up after launch. If delivery risk is blocking launch, let’s talk.

Built an issue tracker cli for Claude and added a force graph view in the terminal.

The forcegraph is mostly not useful except to kind of zone out while Claude works but it looks cool lol.

Now I know enough about AI to grok what MCP is, it irritates me because fundamentally:

Make a cli tool instead! Then a human can use it too.

It’s not stateless? That’s what unix sockets are for!

Once again we’re reinventing stuff from the 70s and making it worse.

I usually make the list and then not look at it until I'm done, it's the only way I won't feel pressured and it lets me keep the just in time discovery ADHD is good at.

Building user interactivity on your website without a backend. - nekomimi AT Proto lets you build social features without running any backend. The user authenticates, you write a record to their repo, Constellation indexes it, and anyone can query those backlinks. The data ...

I love this. atproto at its best nekomimi.leaflet.pub/3m6pcklmvt222

Building a stack visually

1st image, building an app
2nd image, app is published to PDS
3rd image viewer renders it in an iframe and injects the SDK

The hypercard part would be some nice builder UI but the user could also just manually push to their PDS and compatible appviews would discover the app.

By editing the injected SDK the user chooses where their data is saved. One viewer could be pure PDS, one could be some centralized store, one could be local only.

If you trust it there's no reason you can't fetch the "backend" source from the stack definition and run it in an isolated sandbox

Yes!! at-proto needs hypercard. As soon as I get some time I'm planning to build it out.

But each stack/card a record. UI definition is either straight compiled HTML ala ChatGpt Apps, or some schema that defines what the view should render.

AppView injects a standard async SDK into the UI.

Stop treating it's subjective opinions on WHY it did something as important. It's noise. It does not know. Focus on the important shit, is this maintainable? How is this decomposed? How is this going to evolve once I build on it?

Stop treating the AI like a human. It is not. It cannot learn and it is frankly stupid to expect it to. It pattern matches to what you give it. So give it good patterns to match. This isn't hard.

And if you literally cannot find a single way to speed up your work flow with a fucking magic box that spits out 80% correct code, the problem is YOU not the magic box. Learn to use your tools better. Be more creative.

I find so much of the AI discourse tired and obvious. No shit if you one shot things without reviewing the code you'll get shit. No shit you have to manage your context. No shit producing a plan for the ai to follow before you code is going to result in a better build. Maybe I'm just in a bad mood.

I just think this is such a neat functionality for a tamagotchi.

Introducing `steve`: a CLI for driving macOS apps.

Like playwright for browsers it's drivable by code agents. Tell them to "manually test using steve".

→ github.com/mikker/steve

Post image

OpenAI is officially supporting logging in to OpenCode with ChatGPT.

Bullish on OAI, bearish on Anthropic after this.

Enabled by projects like opencode.

So I'm pretty sure the mcp UI stuff is much more powerful than what's being advertised. I see a truly personal, just in time software future very very soon.

That's my next project.

This is the post that made me try opencode. After a day of user holy shit. Seriously good piece of software. I'm excited to build with what the server enables.

today is a great demonstration of why competition is the most important thing in the world

we're not doing "people who criticise genAI don't understand it and are ignoring its utility" any more. get better material

AI makes coding feel closer to art. I'm not sure how to explain it but it does.

I take back any reservations on opencode. I'm a convert.

It's simply a more steerable, consistent experience than Claude code and has so many more extension points.

It's awesome.

Image from X

in opencode v1.1.11

you can now use your ChatGPT Plus/Pro plans in OpenCode

/connect to set it up

Wow, a comment on HN I actually agree with!

A screenshot of the cutaway gag from Family Guy where a man is driving a Hummer and watching Madagascar.

HAHAHAHAHA dude this PR is so fucking LGTM, it makes me wanna merge without looking

I lied. I kind of love it after doing more digging. Way more extendable. It doesn't follow my instructions as well which I'm sure has to do with the system prompt but I can see a lot of cool integrations possible with this.