I was thinking more along the lines of someone who wants to be an open source developer but doesn't know how to do it so will just use AI to vibe code litterally the entire thing.

As someone who's worked with data from the backend of these systems I feel your pain. Why does someone need 10 different DoB's over the course of 3 visits and why did those 3 day visits occur on 8 different days?

As a physical therapist, one thing that bugs me (and I know it’s not a software limitation but spending choices) is having to wedge my notes into a format designed for a physician or something else.

most of the issues with our charting software boils down to dumb shit like that - proprietary systems that are incapable of communicating, or just good ol’ fashioned garbage design

This is everywhere. I deleted a more thought out response when I realized that someone probably handed over ALL of the backend databases of patient data to an LLM (it's HIPAA compliant so don't worry *wink*) so that it can spew out MORE bullshit that doesn't work, but a lot of money was paid for it so we're going to use it.

Oi,I feel you so hard on this, as a software developer who works on medical devices and software used in medical contexts, holy shit the pickings are so bad.

Why can't someone just conjure up solutions to all of this (which is totally doable technically)? Because FDA regulation forces us to follow Byzantine release processes and meet very exacting requirements for the software end product.

I could probably knock together at least two or three of your wishlist items in a week, and it would take 1-3 years for it to make it to the first customer. Even patches and fixes I build for existing products take 3-6 months to make it out the door after they're done.

All of that to say, these make friction points that make it difficult for developers to coordinate and standardize without introducing mountains of extra bullshit that didn't need to actually be in the product or software in the first place

I work for a division of a health tech company and I can say that we hear you and value this kind of feedback. There are folks I work with who have dedicated themselves (including myself) to seek solutions to problems like these, and as a former integrations engineer, some companies are better at interoperability than others, despite the efforts of the HITECH Act in the US. My frustration is time and getting the attention of those who greenlight projects to work on things like these. On one hand, I feel OSS could help solve these problems sooner, but on the other hand, healthcare has some very opinionated people and things could get messy.

have you looked at ersatztv?

Multiple things exist to do this.

I use kodi on a pretty basic level for about 15 years now. I'm almost sure that somethi g like this should exist, no?

I mean, that's sort of already happening: Model Collapse. It's just a matter of time now.

This apparently does exist and is called OMR(Optical Music Recognition) software

If I was to build something like this, could you beta test? Sounds like a fun weekend project for me 😄

This is essentially what I was about to say. I want to be able to scan my sheet music and end up with a perfectly clean, crisp, beautiful copy of it on my tablet.

Please take my money. I would pay thousands for this

There goes 90% of the internet ...

At least for YouTube clickbait there is crowdsourced DeArrow, from the same developer as SponsorBlock. And you can use it in FreeTube!

Honestly, existing alternatives and frontends get surprisingly close.

A software version of taxes!

gotta rely on ex-wives of billionaires instead i guess

Public service announcement that, once split up, this would be a significant but not life-changing amount of money. At least in the West.

Literally just got pissed off enough to give up for a bit searching for a hard drive enclosure. Fucking pages and pages of docks. I typed in the god damn parameters of what I am looking for, stop showing me shit that isn't that. The assholes miscategorizing their shit so it comes up in more searches should be executed.

Working on the second: https://menagerie.circuitforge.tech/snipe

Still early days but the filters and trust scoring work really well

I've used https://kagi.com/ for the past year or so. It's much improved search compared to current Google.

If you run Android, check out "Energize" on fdroid.

Waistline kind of works like that.

And I think there are a few self-hostable solutions out there.

If all you need is calories couldn't you just use any counter app for that? At that point you're just adding numbers in.

Sadly that's a human problem that technology is hard-pressed to solve (if not impossible).

I hear Lemmy is pretty good but I've never used it.

I am afraid that would only be possible if it's faked, with everyone else except you being a LLM

Or/and if almost every real user is shadow banned to other real users except bots

Who would bet against those fires in a platformike that?

I used to work for a startup back in 2009 that used semantic tagging and your social graph to improve results on the big three search engines. It was a pretty cool idea with the semantic tagging (type "apple" and you would get a drop down to select apple the fruit, Apple the tech brand, Apple the record label, etc.) and then your search results were filtered to what you were looking for. They pivoted and added your Facebook network so others using the system that you knew would prioritize results with the assumption that if you were searching for something that someone you knew had already searched for, it was likely that their findings were relevant. Yes, a ball of privacy issues to be sure. Always anonymous, and you could opt out of sensitive searches being included, etc. But, it did provide some means of social proof in your search results. Short lived job for me as funding ended and I moved on to create my own company.

In a way, I think the early AOL had this idea.

Half of it exists. https://yacy.net/

Problem is, index too small. Be the change you want to see in the world, I guess.

I had an idea, personally I just use the Google phone app it seems to do a pretty good job of blocking spam calls somehow. However I had a funny idea. I suspect that these robocallers May be able to recognize the we're sorry your call cannot be completed as dialed tone and words which may signal to them that the number doesn't work.

I've been doing it this way for a while too. Glad it's working for others