We never had a working Siri so we can wait another few years. Not that it matters at all.

Could be punishment for devs who flew through the PRs without care.

Thanks, it's fun to build, and I learned a ton.

Nice! Do you have a set of data points that you get your claw to report on daily?

You can reach out. It's all made with Claude Cowork to extract the Historical data from my 4GB Apple Health export (accessible from the Apple Health app, then dumping it into Lovable to build a dashboard with the data points that I'm interested in, and then a little design pass to make it look nice. You also need to figure out what metrics are important to you and then playing with it. Lovable know "how" to build dashboard, so it's about telling it what you'd like to see/learn about.

I've actually learned a ton by simply requesting the data from Apple via their export in the Apple Health app, and then loading it into Claude Cowork and letting it analyze it. I also have about 4GB of data dating back 2015. You can isolate a few dimensions you're interested in and let it guide you. You can also ask it to export that data only and then dump it into Lovable for example to build a small dashboard. For the longest hikes, that would require regular outputs... So I'd recommend that auto-export app, or, if you save them into Strava, building something to extract from you and your friends' accounts to build a leaderboard. Do it!

Three people building frontier AI: Boris Cherny (Claude Code), Qasar Younis (Applied Intuition), Llion Jones (Transformer paper, Sakana AI), all spent time in Japan and credit it with changing how they think about building.

(you could fix your link so it's clickable) 1. thanks for building this. I will get back on my iron deficiency diet. I now understand it takes over 7 weeks to reliably fix 2. when doing data input, I'm lazy, especially for the blood age calc. So my process is: upload list + my blood results to the LLM and spit out the list of values I need (terrible privacy job right here for me) but anyway, I wonder if you could offer another route for data input, like a text field, with the full list and empty values, that I could copy to an LLM and ask to populate with my results and then spit back to paste into the form. Keep up the good work!

thank you for your kind words and good feedback.

importing labs is something I think about a lot and I think the solution will be something along the lines of what you suggested. Since I need to keep 100% privacy which I publicly promise.

It's really fun! Thank you. On the result screen, let me click on the locations so that I can learn more about them. Some museums I didn't know and would click immediately to learn more about them. Or even add a little explanation of what they are.

Yeah, good idea. love that! Sweet, thanks so much for the suggestion. I've taken a note

Sounds great. Would everyone "accept" the result, or would it be worth adding a little LLM explanation of why the result should content everyone by explaining how the game retains elements or this other voted game and that other voted game, to try and make people go "ok, sure"? I'm not a huge gamer so maybe this is an obvious reaction they would get from their own experience when seeing the result without needing a LLM explanation.

Maybe far down the road. For now I'm fine with a minimal tool that does not aim to take out the human interaction in the group haha. If the tool finds a small community more features might be added like content-based recommendations, but I don't want to drive up the costs right now