思想与社会2023 · 分享于 Hacker News
The Curse of Recursion: Training on generated data makes models forget (2023)
探索《The Curse of Recursion: Training on generated data makes models forget (2023)》的原文与 Hacker News 社区讨论，查看来源并加入阅读清单。
从论文，走向讨论
看看社区如何讨论这个想法。
来自 Hacker News 的原始讨论。不同的视角、疑问，以及值得继续思考的线索。
最多展示 100 条相关讨论。论文可能出现在评论中，并非原帖投稿内容。
在讨论中发现的引用
按 HN 条目顺序展示最近的 50 条引用来源。
Ythinkingemote ·
The Internet Is Full of AI Dogshit
805 赞同587 评论
Ytosh ·
Gemma 2: Improving Open Language Models at a Practical Size [pdf]
328 赞同172 评论
Yzerojames ·
No "Zero-Shot" Without Exponential Data
187 赞同118 评论
Yjacquesm ·
Tell HN: We should snapshot a mostly AI output free version of the web
136 赞同60 评论
Ysurprisetalk ·
The Curse of Recursion: Training on generated data makes models forget (2023)
122 赞同107 评论
Yrntn ·
You Sound Like a Bot
57 赞同50 评论
Ygorbachev ·
Two US lawyers fined for submitting fake court citations from ChatGPT
28 赞同24 评论
YFanaHOVA ·
LLaMA2 isn't “Open Source” – and why it doesn't matter
28 赞同6 评论
Y_delirium ·
Why This AI Moment May Be the Real Deal
3 赞同0 评论
Ymalshe ·
Generative AI Degrades Online Communities
3 赞同1 评论
YYeGoblynQueenne ·
The Curse of Recursion: Training on Generated Data Makes Models Forget
2 赞同0 评论
Yda39a3ee ·
The Darwinian argument for worrying about AI
2 赞同0 评论
Ysizzle ·
Google's Privacy Policy Now Admits to Collecting All Your Data for AI Training
2 赞同0 评论
Yindus ·
The Curse of Recursion: Training on Generated Data Makes Models Forget
1 赞同1 评论
Ygmays ·
The idea maze for AI startups (2015)
1 赞同0 评论
Yjerjerjer ·
ChatGPT will make the web toxic for its successors
1 赞同1 评论
Ywvoch235 ·
Using GPT-4 Vision with Vimium to browse the web
1 赞同0 评论