i write music and songs (kgw.me) and now i guess i write software too. i think we should all be writing more personally...
caught a wild hare? if i already caught it and had it in hand, i wouldn't be off on a chase. anyway, i decided to make enough‡ -- my local weights agent harness for non-technical writers aka normies aka the sort of people who have high powered mac computers but don't ever open terminal -- which up until now has required multiple terminal operations and ran in the browser, into an authorized mac…
one thing i realized about needing to make apps: if your webapp does something cool or useful enough, people will just use that. the thing i realized was that i wanted to use midjourney so bad in 2022 that i installed discord, which i had never used. i hate it to death to this day but i made it work!
or honestly BECAUSE you did
not that i haven't had plenty of thoughts ... being unemployed really makes you doubt the value of thoughts. i wish more people listened to my extremely unusual podcast: e47.net and i wish people would try my software.
had to write this somewhere to avoid spamming my cousin with a strand of "the thing is" text messages. figured maybe others might appreciate it too. so i made a harness; had a bunch of initial motivations since discarded and plenty that persist, but it is definitely an LLM harness. it is designed to be a very knowledge-filled tireless & totally free as in beer sounding board, brainstorming buddy,…
let me yell at you. what does chris walla, the producer and musician, mean when he says "burn ai to the ground?" it's just not an option, any more than "bury war underwater" is. we all want both of these things, but capital will continue to demand them. ok bye cloud
thank heavens i hedged my bets by playing the long game, i.e. trying to come up a piece of software that is useful today but should still be accessible and useful in 20 years. what's the rush? now, the downside is, it can be tough to know if anyone thinks it is useful today. i think i am either more paradigmatically idiosyncratic than i think i am (or would like to be) or, more generously, a bit…
i realized i can just set the repo to public and not really tell anyone until a few people have sanity-checked it. versus trying to get everything exactly right all at once. drizzle the launch across a timeframe. nothing good was ever built in a day (probably). you can check it out here: 0g.software/enough if, you know, you read this somehow. as for how you'd tell me about my mistakes...
it can't ONLY be a lack of confidence in our writing abilities ... or i mean, maybe it can. that's a big attack surface, when you think of all the ways imagining your audience can bite you. as discussed, i just pretend (?) nobody really cares about the mistakes i make or words i misuse as long as i'm earnestly attempting to communicate information that i care about. in practice, this kind of feels…
you just have to do it, i guess. you have to be proven right or wrong. it's definitely good not to have bet the farm on something, or bet anything, really. a few domain names. it's hard to know what people actually want and want to stay away from vs. what "the media" implies they collectively demand & despise. there's plenty to be said for the courage of one's convictions, i guess. and only one…