Oxtak 确实是我的第三家公司,但它是我一路做下来的延续。往前倒一点:从手机设计说起,主要是二十年前的飞利浦手机,我做过一些小众的、非常独特的产品。
Oxtak is indeed my third venture, but it's a continuation of everything I've built. Go back a bit: from mobile-phone design, mainly Philips phones twenty years ago, I worked on niche, very unique products.
我参与过 TAG Heuer(泰格豪雅)Meridiist 的设计团队,那是这个隶属于 LVMH 集团的手表品牌做的一款手机。我做过很多手机,等到智能手机出现,也就是 2007 年的 iPhone 时刻,我就搬到了深圳。之后我创办了 Omate,一款直接连入电信网络的智能手表。
I was on the design team for the TAG Heuer Meridiist, a phone from the watch brand under the LVMH group. I built a lot of phones, and when the smartphone arrived, the iPhone moment in 2007, that's when I moved to Shenzhen. Then I created Omate, a smartwatch connected straight to the telecom network.
关键在于,它是一台独立设备,比 Apple Watch 早了两年。它不是挂在你智能手机上的配件,而 Oxtak 遵循的是完全一样的理念。
The key point was that it was a standalone device, two years before the Apple Watch. It wasn't linked to your smartphone as an accessory, and Oxtak follows the exact same philosophy.
我甚至不叫它独立设备,我叫它side device(旁侧设备),一个在你身边、却真正独立、不与任何东西捆绑的东西。
I don't even call it standalone, I call it a side device, something alongside you but truly independent, not tethered to anything.
你大概见过市面上那些 AI 录音设备。它们大致做三件事:实时翻译、录音、问 AI,而通常这一切都装在你的手机里,要么预装,要么是 App。我的手机本身就有一个很好的录音功能,甚至能做语音转文字和摘要。
You've probably seen the AI recorders out there. They do about three things, live translation, recording, ask-AI, and normally all of that lives inside your phone, built in or as apps. My phone already has a great recorder that can even do speech-to-text and summaries.
后来出现了配件,就是那种贴在手机背面的小卡片,Plaud、Pocket、TicNote。他们把为什么你会想要一台单独设备这件事讲得很好:更方便,而且你的手机不必被录音占用,还能腾出手来。
Then came the accessories, those little cards you stick on the back of your phone, Plaud, Pocket, TicNote. They did a great job explaining why you'd want a separate device: it's more convenient, and your phone stays free instead of being tied up recording.
但问题在于,它其实还是连着的。音频依然得传到你的手机上去处理、存储、生成摘要。翻译设备是这样,连智能眼镜也是这样,音频最终还是流向你的手机。
But here's the thing, it's still linked. The audio still has to be sent to your phone to be processed, stored, and summarized. Same with the translation devices, and the connected glasses, the audio still flows through to your phone.
所以我想要一个真正独立、并且隐私优先的东西。
So I wanted something truly standalone, and privacy-first.
你想想看:此刻这场对话正在被录音,而我根本不知道那段音频存在了哪里,你多半也不知道。对这样一场公开的对话来说,无所谓,你注意自己说什么就行。
Think about it: right now this meeting is being recorded, and I have no real idea where that audio is stored, and neither do you. For a public conversation like this, it doesn't matter, you just watch what you say.
但换成这些小设备的使用场景,那是私下的讨论、团队会议,甚至是战略会议,人们不想再做笔记了。很方便,可那段音频就是一个源头,而有朝一日,它可能被用来对付你。
But with these little devices it's private discussions, team meetings, even strategic meetings, where people don't want to take notes anymore. It's convenient, but that audio is a source, and one day it could be used against you.
我快速解释一下。语音转文字,本身就已经是一层 AI 过滤,音频变成了文字。每一家厂商都会告诉你,它有 95%、也许 96% 的准确率。
Let me explain quickly. Speech-to-text is already an AI filter, audio becomes text. Every vendor will tell you it's ninety-five, maybe ninety-six percent accurate.
可即便它达到 99.99%,你依然可以说,我在会上从没说过那句话。别人会说,得了吧,白纸黑字就在转录稿里。而你可以反问:源头在哪?证据、凭证,是那段音频。
But even when it hits 99.99 percent, you can still say, I never said that in the meeting. People say, come on, it's right there in the transcript. And you can answer: what's the source? The proof, the evidence, is the audio.