这个无损全自动书籍扫描仪后面衍生除了个开源项目

This is way better.

This one is more fashionable👍 Could you provide more detail about this device?

@DevSkg

@ShoryuPaw @DevSkg 谢谢，这个原理是 “真空负压吸附定位扫描 ➔ 气流喷射无接触吹动翻页”，就是价格有点贵，好像差不多10万到20万美金

@DevSkg 贵在翻页方式，也就只有图书馆档案馆做数字化且涉及稀有度和价值极高的书籍，才有可能有应用场景，不是用来扫描原版孤本都对不起这价格

@ShoryuPaw @DevSkg 这个确实超贵

This gives me anxiety about a page getting ripped out

A vacuum system suctions a single page at a time, peeling it across the scanner sensors. A brake-beam sensor to detect if a page fails to turn or if there is a paper jam, at which point the machine stops to prevent damage

BFS-Auto: High Speed Book Scanner at over 250 pages/min

这个看起来非常先进，不过可惜没有量产

@AnthropicAI, you don't have to destroy books to get your "data".

@AnthropicAI This device was used by Google to scan the books in the world's libraries.

Not only does it scan losslessly, it also has the noise removed

Yes, as the book travels over the vacuum slots to turn the pages, the suction actually pulls dirt and debris out of the book. The presenter jokingly refers to this as a "feature" of the machine.

@juancarlosmll Puedes usar una contadora de billetes para comprender pasar hojas.

@juancarlosmll 嗯，原理跟印钞机有点类似。 不过更简单的理解方法是：负压吸盘

@juancarlosmll Cómo las impresoras offset Ryobi que levantan las hojas por vacio

Why did Anthropic destroy all those books then 😡

破壊せずに読み取る過程を正確に自動化するってのが すげー

这好像是把扫描的书页切下来了，那直接去掉背胶扫描不就行了么，甚至都不需要额外的设备，很多稍微大型的扫描仪都可以自动双面扫描。

Can we just have humanoid robots read the books? Seems like it would be faster.

Yes, there it is

失敗したらページが破れそうだけどいいね👍 結局、全ての知識をOCRすることがAIに対して一般的な人類ができる最初で最後のご奉仕。

硅基生物终将统治世界😃

First 100 prototypes were fully automatic page shredders

真相了，👍哈哈哈

I've been meaning to build one of those, might need to start on that soon

does this hurt the animal?

No, it helps save animals. When a book is scanned into an ebook, more people can reaadit, which reduces the demand for the paper materials. This in turn reduces the number of trees cut down, benefiting animals and ther environment.

The fact this exists and most people have never heard of it 💀

It's not lossless.

yes, it's an auto-translation misinterpreted. what's I mean is: non-demanging to the books .

also great for slicing pastrami

🍻

这才是LLM的bootloader吧🤣

Google 20 年前就开搞了，但是现在 Gemini 竟然还很拉💩

肯定没用你的这个产品，所以这么拉

😂这个就是 Google 当初设计的

😅

We’re calling OCR “lossless” now?

@grok how does this machine work?

Is that a @MieleUSA powering the whole thing?

如果后面再加一个焚烧炉，猜猜是哪家公司的产品

🥰

very smart

nice!

这个有超10年了

Google 的开源是2012年

This looks cool but I'm sorry to tell you that analog -> digital cannot be lossless

Beautiful but not usable for old books they are very fragiles ....

@BrianRoemmele

Did anyone else witness that?

Feed the AI god

Can it cut cheese?