Howe

@plantegg 11labs，有好几款音色甚合我心

有点意思。其实我最原始的目的是想要一个晚上播放英语故事的东西，而且我要播放《哈利·波特》。 去买的话，每一本都需要 99 块钱。 我现在的方案是先去下载英文版的 EPUB格式。下载完了之后，让大模型去帮我找一个文字转语音的 AI模型，然后帮我进行转换。 刚刚试了一下，效果还挺好的，然后是转为 MP3 格式。 好像现在手机上，MP3 格式的播放器是不是都需要单独下载？现在大家都习惯用网易云音乐、喜马拉雅之类的app来听，这些似乎并不算是播放器。

But the way, I sincerely note that: The pursuit of perfection is endless, but what matters is the effect of the phased inference. I.e. You will find HTML is not even the final either if you are obsessed and proceed.

I feel there’s no true or false on what method to choose. It just depends on your initial need. From my perspective, MD is still my preference since the content matters to me essentially.

Undoubtedly HTML holds some advantages which MD could never touch, but I have to ask you guys “what’s your primary intention?” I bet MD goes well with the answer of “content edition”(lo-fi), while HTML serves “high fidelity”(hi-fi).

@xicilion 请问应用场景的起点和终点是什么呢？ 将 SVG 信息与模型进行传递，以便交流，然后进一步推进项目吗

@nikitabier @Chilearmy123 Detection is on agenda, how about tort action then?

@Chilearmy123 Content theft is a losing game. We already have duplicate detection for videos working in internal testing. Impressions & engagements on reposts will accrue to the original author.

Being a clipper in 2026 is so much better than a creator/KOL You make WAY more money You can infinitely scale (hire people under you, make more accounts) And the biggest bottleneck (make more content) doesn’t exist as there’s unlimited content If I was starting today I never would’ve been a “creator”, I would’ve just been a clipper I think platforms will eventually push back against it as original creators grow frustrated with the rise in clipping and they can’t compete But not stopping anytime soon

@changwei1006 检索只用Everything Explorer只用来预览，而图片浏览自然是一坨🥲

在网上搜索explorer.exe的替代品，搜到一个大家都在推荐的files.community，结果用了才发现这玩意儿比Win11自带的explorer都还要更卡，右键菜单五秒才出来 UI也没有比Win11好看多少 打开一个包含有上千张照片的文件夹直接卡死，Win10和Win7自带的资源管理器都没这么差 那我换了个寂寞？？？

@skaas777 等同：没有鼠标的笔记本，点击关闭了触摸板控制

antigravity真特么是个天才，让它去运维服务器，它帮我把密码登录给关了，现在我也进不去服务器了。

I saw this and created that.

@siantgirl 这已经不能算主持人了，只能叫做阅读器 原本有温度的内容传播，随着边缘化和式微，已经沦为白噪音了

湖南经视频道已经让AI主持新闻节目了，画面也是AI切换，全面AI化了。感觉播的新闻都是AI合成的，太魔幻了。

From third to first in ONE MOVE⚡⚔️ Valentin Debise wins Race 1 ahead of @AlbertArenas75 and Can Oncu🔥 #HungarianWorldSBK 🇭🇺

Nowadays, I couldn’t even distinguish agent from agent where the former is a customer service and the latter is a bot.

No wonder it’s saying “Weather Unavailable” just now, which I thought it literally a regular stroke🥲

Manus收购还是被叫停了 “外商投资安全审查工作机制办公室（国家发展改革委）依法依规对外资收购Manus项目作出禁止投资决定，要求当事人撤销该收购交易。”

“这不可避免要让低质量用户抢占模型训练的集群”，可以展开讲讲吗？

@sama @gabeeegoooh Though the story suits kids well. But the new model could really close the logical loop.

Here is a manga made by ChatGPT Images 2.0 of @gabeeegoooh and me looking for more GPUs:

Really excited for this week! Next up, we've got something to show you at 12 pm PT today.

@lisaawrites 菩提本无树，明镜亦非台 本来无一物，何处惹尘埃 Bodhi is no tree, The bright mirror no stand. Nothing exists from the start, Where can dust land?

According to Philosophy, the highest form of peace is to have zero desire to be understood, admired, pitied or even known.

@icanvardar Absolutely. Some logic couldn’t be figured out unless Figma at first

figma isn’t dead. you really think llms can match the creativity of a great designer lol

@mweinbach All I get is that, the most improvement from ChatGPT is crossing platform which could free from the app wall and execute actions with more plain methods

This is why it sucks and AI replacing UI and apps makes no sense The order flow of using the Starbucks app (19 seconds) vs ChatGPT (103 seconds)