153
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 02 Sep 2024
153 points (100.0% liked)
TechTakes
1489 readers
29 users here now
Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.
This is not debate club. Unless it’s amusing debate.
For actually-good tech, you want our NotAwfulTech community
founded 2 years ago
MODERATORS
Also I'd hate to see Scrivener touch AI - esp because they sponsor nanowrimo and still seem connected https://web.archive.org/web/20240902130810/https://www.literatureandlatte.com/nanowrimo
Scrivener is a hero product in my research/writing as an example of a software product that is designed for concrete purpose
A while back one of their reps did say somewhere on Reddit that they have no intention of adding any LLM features to Scrivener. Granted, they said that in the context of moving towards a subscription model and talking about features that don't work with their current business model, but still. Unless something has changed recently, they seem to want to stick to being a one-time purchase without any cloud-based services whatsoever, including AI, for their next major version too.
I hope so. I know the founder designed it and then learned how to code to build it himself. Hopefully he's still running the show and he's a good one
that is rad as hell
here's the interview where I learned that. https://medium.com/@timetolose/my-interview-with-keith-blount-a49c7764f26f
Official word from Scrivener here https://xcancel.com/ScrivenerApp/status/1830556231431254328
I use NovelAI myself. But you gotta provide good context since it mimics your own writing and isn't an instruct model. It's more of a "yeah, and—" for brief passages.
read the fucking room before you come in here and advocate for your favorite plagiarism machine
I'll step up and say, I think this is fine, and I support your use. I get it. I think that there are valid use cases for AI where the unethical labor practices become unnecessary, and where ultimately the work still starts and ends with you.
In a world, maybe not too far in the future, where copyright law is strengthened, where artist and writer consent is respected, and it becomes cheap and easy to use a smaller model trained on licensed data and your own inputs, I can definitely see how a contextual autocomplete that follows your style and makes suggestions is totally useful and ethical.
But i understand people's visceral reaction to the current world. I'd say, it's ok to stay your course.