_neil

Surely grok has a built-in market with too-online, retired boomers. It's free real estate.

If you haven't already, set your vault directory in iCloud to `Keep Downloaded`. I had issues with notes going missing or just being empty shortly after saving, and this setting fixed that.

But is that really a market worth going after?

I had this happen a bunch when I was using iCloud sync on multiple devices. I think it was mostly solved by setting the directory to “keep downloaded” (right click on it in finder and it’s the second option).

That said, I’ve switched one vault to git and have had no issues there.

Direct link to pz for those on mobile: https://github.com/joelreymont/pz

I'm building a Home vault like this currently. With the new obsidian cli, you can do a lot more with letting an agent manage things and update dashboards, etc. for you.

Much better bangers about fraud:

Dead Prez - Hell yeah: https://youtu.be/kGjSq4HqP9Y?si=_z6jb0Vfo7_PiITQ&t=82

Maxo Kream - 5200: https://www.youtube.com/watch?v=7kC9j6Zp-kg

Maxo was actually arrested for racketeering, though not due to this song specifically (I don't think).

Woefully and eternally under-maintained but

https://descend.org

damn right

Agreed. Zed is my daily driver and I love it, but the autocomplete is not good. I end up disabling AI altogether.

Ah you’re right. Pretty cool.

Might have been audio?

https://phys.org/news/2018-02-power-grid-fluctuations-hidden...

The current version does not. I’m not sure if the official Health export includes that info. My current doctor unfortunately uses some app that doesn’t integrate with Apple Health. It would be a great feature here though.

Awesome! I was just looking at dxt. I’ll take a look tomorrow.

No, you can ask arbitrary queries. I’ve found that foundation models (and Claude in particular) do a great job at writing sql. It even seems to handle time series questions pretty well.

Good idea. The video on Twitter has a few different queries (though I don’t expand all of them) in order to build the dashboard. If you pause it at different points, you can see some of the questions it’s asking and then subsequently writing queries for.

All that said, I’ll try to add some examples to the readme.

Very cool. I already had this started at GitHub so I must have seen it in the past.

I love the use of a notebook as an updating dashboard here. One of my demos before this MCP server was an evidence.dev dashboard.

It seems like you still need to export the database here, so maybe same amount of steps since duckdb can import CSV files directly as well.

I think supporting the native Apple Health export feature would be a good idea. It just takes longer than the CSV export, so I started there. I would imagine the native export has more data, though, which might be nice.

Agreed. I'm fine jumping through a couple hoops to get my own data in this case.

I haven't gone deep on prompting around health data specifically, but I've built some other data agents. Agree that there isn't always an easy path, depending on the domain. It might be worth exploring multiple subagents with different focuses. I.e., have one subagent that is going to focus on writing window function queries in order to find interval variations.

My general belief around data agents is to offload as much as possible to SQL.

Great minds :)

I haven't experimented with MCP too much because I have some reservations about it, but I decided to go MCP-first for this to see how it feels to prototype around it. My typical flow would have been to go sqlite+sveltekit.

Simple Health Export CSV was just the fastest route to proving to myself that the idea would work. I really considered making an app but decided that was one nerd snipe too many for the given task.

But I'd like to revisit it because it would be a natural fit to just put the personal trainer flow directly in an app.

It would work with those devices since it all funnels into apple health. I have a Withings scale as well so it’s part of my export. But you’re right, there may be easier integration with Oura.

I actually started with this but the export always had the appearance of stalling unless I left it for long periods. For a first version, the CSV export app was much quicker. And actually, the author of the app (Eric Wolter) has a nice site that will convert the xml format to csv using the FileReader api, etc.

I would like to go back to this though and add it as an integration option.

My experience is that you can be very targeted in your promoting with Claude code and it mostly gets good results. You can also ask it early on to create a branch and create logical commits as it works. This way, you can examine smaller code changes later in a PR (or git log).

Or if you want to work more manually, you could do the same but not allow full access to git commit. That way it will request access each time it’s ready to commit and you can take that time to review diffs.

The community for GSAP is still amazing and there’s years of forum posts for most questions you might have. Motion and anime are great but I’m skeptical that either have a lion share relative to GSAP.

A draft model is something that you would explicitly enable. It uses a smaller model to speculatively generate next tokens, in theory speeding up generation.

Here’s the LM Studio docs on it: https://lmstudio.ai/docs/app/advanced/speculative-decoding

Great point.

Not on-device but for local network I’ve been looking at Speaches[0]. Haven’t tried it yet, but I have been running kokoru-web[1] and the quality and speed is really good.

[0] https://speaches.ai/ [1] https://huggingface.co/spaces/Xenova/kokoro-web

It’s not out yet, but the upcoming Framework desktop [0] is supposed to have a similar unified memory setup.

[0] https://frame.work/desktop