©CC BY-NC-SA 4.0

频道: cx.ms/channel
笔记: cx.ms/memo
博客: cx.ms/blog
剪贴: cx.ms/clip
社交: cx.ms/sns
#article #read

In reply to nevent1q…hvxn
_________________________

美国版权局不认为使用版权材料训练 AI 是合理使用,特朗普随即解雇了局长

美国版权局上周五发表了一份有关使用版权材料训练 AI 模型是否是合理使用的报告,次日特朗普下令解雇了该局局长 Shira Perlmutter。科技巨头和 AI 公司是特朗普的支持者,其中马斯克(Elon Musk)旗下的 AI 公司是众多希望使用版权材料训练 AI 模型的公司之一。美国版权局的报告区分了研究用的 AI 模型和商业性 AI 模型,认为研究用的 AI 模型能带来变革,但商业性模型生成的是与训练使用的版权作品相似的输出,不太可能被视为带来变革性。商业用的 AI 训练类似人类学习。版权局更多是站在了版权所有者而不是 AI 公司这一边。

- https://www.businessinsider.com/ai-training-copyright-laws-big-tech-fair-use-openai-meta-2025-5
- https://www.solidot.org/story?sid=81263

#Copyright #AI

via Nostr@cxplay_clip
#吐槽

In reply to nevent1q…8m95
_________________________

已读回执是社交里面的两难选择. 应该把电子邮件当作普通的纸质邮件的通讯形式来使用, 如果要追求回执就不应该用邮件, 对面要求用电子邮件就说明对面也不希望给你回执, 任何时候都不会主动给回执.

via Nostr@cxplay
#吐槽

我开始怀疑哔哩哔哩是怎么盈利的了

via Nostr@cxplay
#吐槽

今天为了分析 Nostr 的异常客户端让 AI 写了一个简单的 WebSocket tarpit 服务器, 然后最后发现是自己人, 人都麻了, 这就是不好好沟通的下场.
而这个 tarpit 到底应该怎么翻译呢, 焦油坑? 沥青坑?
好是好玩, 但还是要占用连接数和自己的硬件资源, 观察完还是用 WAF 屏蔽掉算了.

via Nostr@cxplay
#吐槽

「韦伯-费希纳定理 ... 人的感觉是和实际的物理量的对数成正比 ...
按照这个规律, 假设你活到八十岁, 那你生命的中点当然是四十岁.
但假设你四岁开始有记忆, 在你的对数感觉里, 生命的中点其实是十八岁.」

「本福特定律 ... 这些符合本福特定律的数据有一个共同的特点: 大小跨越好几个数量级, 但 log 一下就比较均匀 ... 接近对数均匀分布. 在数学上 ... 对数均匀分布的数据集就是会符合本福特定律, 在自然过程中随机演变而来的数据很多都 ... 会不由自主地具有对数尺度均匀的特征, 所以符合本福特定律.」

【毕导】这个定律,预言了你的人生进度条_哔哩哔哩_bilibili
https://www.bilibili.com/video/av114478410699738

via Nostr@cxplay
#吐槽

Dear Nostr client developers, relay automation software developers, protocol extension software developers, please add proper User-Agent headers to your HTTP clients and WebSocket clients, and if necessary declare the client's characteristics and IP range in both directions.

Last night, I blocked some massive resource abusers for my community relay. They all had one thing in common: without a User-Agent, I had no way of knowing who was behind the IP. I recommend that all client developers include the full and correct User-Agent header for both HTTP clients and WebSocket clients, especially for unsolicited messages, and preferably with contact information. I realize that this header can be easily tampered with, but as long as the administrator of a particular dedicated facility declares the IP range and User-Agent characteristics of their client in both directions, this is still valid, as is the case with Googlebot crawlers.

Since I had no way of knowing who was who, I made the mistake of blocking the bot used by Mostr.pub to check the validity of NIP-05 identity in user profiles for a few days, which led to a few minor problems. It was only after a targeted check of the server logs that I realized it was hiding in a visitor with the User-Agent "Deno". cc: Alex Gleason (npub1q3s…d26p)

via Nostr@cxplay
Back to Top