I built clawr.ing so my agent can call me on the phone for urgent things like emails I’m waiting for or issues in production for any of my products

Comments URL: https://news.ycombinator.com/item?id=47237322

Points: 1

# Comments: 0

I thought it would be cool to get real phone calls from my AI agent letting me know about urgent things I care about, so I went about creating clawr.ing.

Openclaw has a voice call plugin but it needs a lot of setup with external services, API keys, and webhooks, and it doesn't let you do things like interrupt the agent or have it use tools during the call while you listen to hold music. I also wanted to build something super easy to set up where you only have to copy paste a single prompt.

You can tell it to call you when something happens, like to watch for an important email or monitor a stock price. It works well with the heartbeat functionality that's already included.

I'm based in Portugal and it calls my number fine so you can get your openclaw agent to call you anywhere in the world.

The service is already above $100 MRR with 20+ subscribers doing calls daily.

Your feedback is super appreciated!

Comments URL: https://news.ycombinator.com/item?id=47234485

Points: 4

# Comments: 1

Comments URL: https://news.ycombinator.com/item?id=47221783

Points: 2

# Comments: 0

I thought it would be cool to have it reach me proactively when something I care about happens instead of having to look for a notification or ask it directly

The result is pretty surreal, being out for a workout or groceries or something and getting a call about a stock price or an important email feels like I was suddenly transported into the future

It took quite a bit of effort to set up though, there’s a lot of complexity to routing calls in a cost efficient manner and generating realistic human like speech

I’m happy with the result though, you can reply back and the agent can run any tool call while it’s on the call with you, and if it takes a bit of time as in for a web search or so it will put you on hold and you will hear hold music for a few moments

I'm already using it myself for all of my ongoing projects, but it would be great to have feedback from external users that would be interested in trying out a solid, albeit not yet that popular library.

It's hard to compete with established libraries backed by corporations on breadth of features, documentation and examples, but I think it's very much possible to compete on quality.

I think the main impediment to adoption is the perception of a library as a toy project, so perhaps I should invest the time into making it look more serious - but doing things like setting up a website for documentation seem like they don't really add much value, as github already has quite a nice interface for documenting a codebase.

Interestingly, this does seem like a similar problem to startups finding their first customers, to whom the advice is usually doing things that don't scale - although that is a bit of a generic thing to say, as it encompasses almost anything.

How would you guys approach finding early adopters?

Comments URL: https://news.ycombinator.com/item?id=44607673

Points: 3

# Comments: 0

but crucially, it assumes that variations in submissions and voting activity are strongly correlated, which I imagine is not necessarily the case

which suggests a different concept of time for ranking posts

the main issue trying to be solved is that currently there is an incentive to find the best time to post something, as the variance in activity causes some times to make for a higher likelihood of post popularity

the suggestion is thus for there to be a different kind of time "tick" - instead of time itself, maybe a site-wide counter of activity such as number of views, votes, or submissions

I find the idea of removing the incentive to find the best time to post interesting. I think an ideal ranking algorithm drives the correlation between the time of posting to the number of impressions on a post to zero.

Interested in hearing y'alls thoughts!

Comments URL: https://news.ycombinator.com/item?id=44499991

Points: 3

# Comments: 6

I think the current ORM model is too restrictive and complex, so I set out to design the perfect minimal ORM for developers that want to ship fast instead of reading documentation and writing SQL migrations that have to run in a world-stopping fashion.

The library is still pre-v1 as I want to add a lot more examples, particularly for React - as I'm now using it on a react project - but the core logic is locked down, and the main motto is that it will never go past 100 lines of zero-dependency code.

Any feedback is appreciated. I really enjoy using it in my own projects and believe it represents a new storage paradigm that prioritizes simplicity and speed of development over micro optimizations.

There are definitely some tradeoffs I would say, but I believe tinyORM sits in a very advantageous position in the tradeoff space - it trades a little optimization for huge gains in simplicity.

Thank you for taking a look! Happy to answer any questions. Your feedback will result in material changes to the library, so please don't hesitate to share your thoughts!

Comments URL: https://news.ycombinator.com/item?id=44499668

Points: 3

# Comments: 0

Having to "if err != nil" every single function call is a big put off - imagine having to "try catch" everything in a language like C#!

Not requiring a world around the brain does not make it more likely, but less, as the probability of brains given that worlds exist multiplied by the probability of worlds is still much higher than the probability of brains regardless of the existence of worlds