"It's really good at summarising code I don't understand".
Uncategorized
1
Posts
1
Posters
0
Views
-
"It's really good at summarising code I don't understand".
But, weirdly, it's pretty bad at summarising code I *do* understand.
-
undefined oblomov@sociale.network shared this topic on
Gli ultimi otto messaggi ricevuti dalla Federazione
-
@anton @thezerobit this project has taken up the majority of my free time and spare thought this year. I have no regrets. It is magic. It has been very rewarding and I highly recommend making visual programming languages.
-
@bryan but, I have a severely negative reaction to companies requiring me to talk to people (or worse, an AI) on the phone. And T-Mobile got to the point where literally everything required a phone call, even things the "T-Life" app or website is supposed to be able to do, that never actually worked.
Even once I was talking to a real person, getting a transfer PIN sucked. Just a miserable support experience.
Not because the people are bad, but because the systems are designed to be hostile.
-
@anton @thezerobit my recommendation is to find a 2D drawing API that you like, and experiment with drawing the basic primitives of your language, and seeing if it scales up acceptably. Once you have that, you can work out how to make those experiments parametric, and how to connect them to the underlying runtime of your language. Pick interaction semantics that are easy to implement. Every button in mollytime triggers on mouse down instead of up for example.
-
@Gina I think the worst part about that is not the (stupid) effect; but the fact some humans in fragile state of mind can start empathize with bots actually start believing they are having a real conversation. There is absolutely no reason to try and make a chatbot look like a human, yet they are still doing it.
And it (literally) kills people. I will never understand how they can keep up with that crap.
-
@anton @thezerobit it also helps that the language runtime itself under the visual representation was something I already knew how to build, as this project follows many prototypes, and other language experiments before it
-
@bryan it was the best deal at the time, and every time I looked into it for several years, for the things I needed. For a dozen years or so, I traveled full-time in a motorhome, so I needed good, fast, unlimited hotspot usage on my phone, which T-Mobile came the closest to most of the time. I usually kept a second plan in addition to the primary phone, either from Sprint or Verizon for times/places I couldn't get T-Mobile service. By the end, TMo had a better network than anybody.
-
@anton @thezerobit and by far the biggest time sink by a long shot has been fighting with Linux's utterly slapdash support for multitouch screens.
-
@anton @thezerobit in general, I've prioritized ease of development over long term correctness, and through the benefit of experience I've mostly avoided painting myself into a corner architecturally. This has allowed me to build out all of this in about 7 months. I hired a friend to help out for about four weeks, and in that time he set up a build system and replaced pygame with SDL3 for me.
