Search
7 results for “mnl”
-
-
If you think your hobby is ruined because someone else does it better with less effort, or someone else does it worse with less effort, is it really a hobby?
Or is it because of the status linked to the effort and outcome?
I’m genuinely puzzled by the virulence around all this stuff. I love drawing and painting, I love making music, I love programming. I poured thousands if not tens of thousands of hours into each of those. In none of those cases do I feel resentment or anger or sadness because either a machine does it better (programming), or millions of people produce some usually garish stuff I can easily ignore (maybe I’m just too offline / my communities are too curated, but I rarely see ai art if not looking for it).
In fact, while I’m quite angry at the companies behind the models, someone pooping out an image with midjourney and being proud of it, that’s awesome. I was proud of my drawings/music/code when they were ass, and I’m sure I got on peoples nerves showing it off.
Maybe they say they “drew” it, then that’s a starting point showing why I like drawing (as in, the act itself) and they might like it too, or just engage them about the images they like and what inspirations they have and then point out (with “critique tact”) why it looks … not great, or just showing images I do think are awesome. Art is quite universal, there’s a good chance common ground here is very close.
Like, generative machines that “are better” than humans is nothing new in music circles. You had “poop music machines” since forever (player piano anyone?). I started with rebirth and ableton and was making some proper poop. And then I got better, because I had people who took me along. If all I encountered was people telling me to fuck off that’s not real music you have to learn how to play guitar, then yeah I wouldn’t have stuck with it for decades.
I totally get not wanting to listen to shitty sample mashups as a guitarist. But then I would just… not listen to it? Not get upset over it, there’s guitar to be played.
-
Well this was an awesome book, highly recommend, about differences in philosophy around authenticity and copies in art (and products, in the last chapter) between Europe and China.
-
RE: https://mastodon.boston/@MSPCAAngell/117485786390076909
Mspca angell is where we got all our animals! They’re on mastodon!
-
I’m gonna play my favourite game of getting yelled at for saying something that will become true in 6 months to a year (since things are accelerating): operating systems as we know them are dead (as they should), most devices will have about 800 different firmwares, and any kind of discussion around what to do with the kernel or what to contribute to in terms of software (regardless of licensing or opensource or proprietary) will turn into “where can I find reasonably well documented source material that has been validated well enough (by humans or otherwise)”. The value of the Linux kernel will be more about the massive amount of decently shown to work drivers than it being a coherent (I’ll leave you to your own opinion on this one) operating system.
One of the reasons we have such massive software complexity problems is that it used to be less costly to add some layer on top of something existing, wether it was good or not, than to sit down, think, and then do the extremely onerous work of rewriting / rearchitecting while shipping something coherent. I’ll just point at termcap and tty handling, or bash whitespace interpolation as an example of the insanity. How many millennia of human (and now gpu) time have been wasted on that nonsense?
Nothing is becoming clearer to me as I replace complex stack after complex stack with a bunch of JavaScript and a few compiler passes.
I’ll leave you with a few examples, using the wasm version of the artcraft libraries. This took me 3% weekly on my Claude and I can now prompt basically any Photoshop/indesign/lightroom/illustrator/excel-y thing I want.
You might think that code disappearing is sad, but you know what’s really hard? Figuring out formal languages to say, “build professional photoshop UI experiences that run on a Nintendo DS style device”, or “what’s a good formal language to build networking stacks”. These are unsolved problems because there’s at most what, like 100 networking stack implementations of reasonable quality out there? We don’t know how to build networking stacks. No person has ever built 50 by themselves to really understand the problem.
We are like the blind trying to scope out an elephant when it comes to software.
-
Soon Linux will just be a repository of reasonably validated datasheets 😭😂
