cleverbit

It’s been fun thinking about what a post-iPhone world might look like—especially in light of Jony Ive and Sam Altman’s recent $6.5B bet on ambient, AI-first devices (via the acquisition of io).

Would love feedback or ideas for where to take it next.

I wrote up my developer journey in a longer blog post including some background on how the project started as a vibe-code experiment: https://richarddas.com/blog/chatgpt-client-for-apple-watch/

lol glad you got the joke

lol it’s me. But thanks for confusing me with an LLM? I guess? Should I complete a captcha for you? :/

Hey thanks for your feedback, it’s really appreciated!

Since there’s interest, I’ll publish the source code to help others learn, and put the app on the App Store for those who just want to use it.

Regarding burpees and other types of motion, I’m looking into their motion characteristics, so that I can include them for detection.

Burpees would likely require recognizing different phases (the jump, the stand, the push, etc) and then when it detects them in sequence, incrementing the count. Something like that, but stay tuned! :)

WOW Murph Workouts! Respect. That is an awesome mission. I’d love to help.

That’s true, and since it’s only tracking head movement at the moment, some further refinements would be needed to account for good positioning. I’d like to develop it further to see what else this technique can recognize.

woah!

It’s called a gym buddy ;) Only half-joking. I found having someone spot you, count or just stay accountable is a game changer. Whether thats a friend, or a personal trainer, it's always more fun in company!

Oh awesome!! I was curious whether it was feasible to distill that kind of information from the Watch. Looking at this, apparently it is! Thanks for sharing.

Garmins are amazing! They are pretty much the de facto equipment for fitness.

I hear you! Finding an app that does just what you want can feel pretty overwhelming sometimes. I like how you phrase “counts reps of any repeating motion”, that’s sparked some ideas! Thanks!

The amount of engineering that goes into these things is truly phenomenal, as highlighted by the CT scans done by Lumafield - check it out here: https://youtu.be/kRpIg23qGY8?si=r2t8_8VHVCwkDDLP

Will do, and thanks for the tip! The more you know! ;)

Love it! Yeah great idea!

I’ll see what I can do, thanks for the feedback!

Simple is the hardest thing. I’ll look into Strava Sync and see what I can do!

Thanks! I’ll post it soon, stay tuned!

I was curious to see if there’d be interest, but I’ll open source the code and release the app soon!

thanks

That’s a great idea. HealthKit doesn’t track anything at the granularity of the number of reps, and while it doesn’t have a workout type specifically for pushups, this could probably be classified as Strength Training. I’ll add it!

Hey folks, just dropping in to share a python tool I put together for working with the OpenAI API. It’s a command line tool that helps with managing assistants, threads, and files.

This was an early attempt at learning python for me, and it was really fun working with GPT to help outline what I wanted the tool to do. It was a pretty big learning curve — I have loads of experience in other programming languages, but less in Python, so it was so cool to just dive right in.

Here’s what it can do:

- List Assistants: See all your assistants ids quickly.

- Handle Files: List, upload, and delete files easily.

- Smart Thread Tracking: I've added a key feature to the tool – it now automatically keeps track of any thread IDs created. This is especially handy since, as some of you might know, OpenAI doesn't offer a native 'List Threads' feature due to privacy concerns. Threads in OpenAI are tied to organizations, and listing all threads could potentially expose private conversations.

One heads-up: You’ll need to set up your Python environment to use it, but the OpenAI docs cover that part.

Thought this might make things a bit easier for anyone juggling with the API. Give it a whirl and let me know how it goes!