@cwebber I really had my eyes opened while on PTO for the holidays, when I used Cursor to configure my Home Assistant platform.
Up until then I was struggling to understand the vast amount of yaml configuration needed.
I have lights, temp sensors, and appliances all hooked in, but the automations sucked. A few days of working through the auto option in Cursor, and I had everything set up the way I wanted and no cloud based megacorp can take that away from me.
@rickpelletier @basetwojesus Seems apropos...
They recently moved a bunch of models to MAX mode, causing heavy users at our company to gobble all their tokens in days.
Guess the low cost days are over, and AI vendors are going to start squeezing the addicts to pay for their addictions.
@hendric @basetwojesus I think the point Masnick is trying to make is that using these tools for personal projects is a way to free ourselves from Big Tech - whether you do it today with a double-cost but super easy tool like Cursor or with a home built custom thing.
Either way, we're heading back into a world where people can just build the software they want
@basetwojesus Yeah, maybe, but is it better to only have systems that we understand?
I'd rather have stuff that does what I want, even if I don't understand everything about how it works, than have nothing
@basetwojesus fair point. I think Masnick and I are both in the position of already knowing enough about the tech to describe what we want competently without worrying that the implementation will be magic spaghetti.
Right now, though, I think the vast majority of 'normal' people think they can't do anything because the barrier to entry is high.
With a 20 bucks sub to Cursor and some kind of computer to run it on, anyone can build themselves a task management app in a couple of days
@basetwojesus my very first real piece of software was built for the IT department at the University of Miami. It was a web page where we had hard coded the area restaurants we liked and categorized by type. Then, it had a little database.
The purpose was to randomly select the lunch choice for the department, with logic to prevent choosing the same type two days in a row, and weights for preferring some more generally popular types.
@basetwojesus it took me like two weeks because I was learning early early .net at the time.
Today, that app can be built by literally anyone in 15 minutes, and nobody has to use some Premium Yelp random picker thing filled with ads