我已经有一段时间没有进行技术更新了，所以我认为现在是个好时机。人工智能技术的发展一直在以惊人的速度发展。有时感觉我们正在迈向奇点级别。I haven’t done a tech update in a while, so I figured now’s a good time. Developments in AI technology have been blazing along at breakneck speeds. It almost feels like we’re hitting Singularity-level strides, at times.
我想总结一些最“令人兴奋”或最令人惊讶的进展，然后假设事情的发展方向。I wanted to round up some of the most ‘exciting’ or surprising developments, then hypothesize where things are headed.
首先让我们熟悉相关术语。AR（即增强现实）类似于 MR（混合现实），涉及将虚拟对象与真实环境融合在一起，而 VR（虚拟现实）则完全设置在耳机内的虚拟世界中。有各种各样的企业口号副产品，如扩展现实等，但坚持基本原则就足够了。为了简单起见，我们不会吹毛求疵，只提及 AR 的总称。First let’s verse ourselves with the pertinent terminology. AR, or Augmented Reality, is similar to MR, Mixed Reality, and involves blending of virtual objects with your real surroundings—as opposed to VR (Virtual Reality) which is set completely in a virtual world inside your headset. There are various corporate catchphrase spinoffs like Extended Reality, etc., but it suffices to stick to the basics. For the sake of simplicity, we won’t split hairs and will just refer to the blanket term of AR.
增强现实Augmented Reality
首先，在增强现实领域，处理能力和新传感器的进步使得对周围环境的实时跟踪变得更加可靠。这催生了即将推出的新功能，并已被 Quest 3 等最新 VR 耳机所利用。Firstly, in the field of augmented reality, the advance of processing power and new sensors has allowed real-time tracking of the surroundings to become far more reliable. This has resulted in new upcoming capabilities, already being utilized by latest VR headsets like Quest 3.
这些功能包括一些平庸的功能，例如在大型 AR“屏幕”上观看电影或玩视频游戏，这些屏幕可以放置在您自己家中/周围环境的任何地方：Such capabilities include the banal, like watching films or playing videogames on large AR ‘screens’ which can be placed anywhere in your own home/surroundings:
可以在自己家里的实时环境中玩的视频游戏：Video games which can be played in a real-time environment of your own home:
或者更多互动类型的视频游戏，可以完全利用您环境中的独特物体和障碍。在这里，我的世界可以在你自己的后院玩：Or more interactive types of videogames which can outright utilize the unique objects and obstacles of your environment. Here Minecraft can be played in your own backyard:
别担心，事情会变得更加Don’t worry, things get far more 有趣interesting……但我们首先要放弃那些更普通和娱乐性的事情。这些新的 AR 环境被展示为允许您在观看节目和数字内容的同时做家务，并使用可以锁定的浮动屏幕：…but let’s first dispense with the more ordinary and recreational. These new AR environments are being showcased as allowing you to do household chores while watching shows and digital content with a floating screen you can lock in place:
但这项技术开始更多地涉及到 AR 和智能眼镜技术与“智能家居”和“物联网”理念的结合，我曾多次写过这一点。这允许用户在基本层面上与自己的联网家用电器进行交互，使他们能够通过 AR 叠加层打开或关闭它们：But where this technology begins to get more involved is the combination of AR and Smart Glasses tech with the idea of a ‘Smart Home’ and the ‘Internet of Things’, which I wrote about several times. This allows users to interact with their own networked home appliances—at the basic level—giving them the ability to turn them on or off via AR overlays:
But a quick disclaimer, some will recall me lambasting these very developments in articles 但快速免责声明，有些人会记得我在像这样的like this one文章中痛斥了这些发展。.
我在这里对它们的报道并不一定等于对这些发展的支持。充其量，我仍然是不可知论者，倾向于怀疑，有时甚至是彻底的敌意——尽管我不一定是一个彻底的勒德分子。我的立场很简单，这些发展My reporting on them here does not necessarily equate to support of these developments. At best, I remain agnostic with a lean toward the suspicious and sometimes outright hostile—though I’m not necessarily a total Luddite. My stance is simply that these developments 是are不可阻挡的，所以最好至少理解它们，甚至可能学会在某些情况下利用它们或与它们共存，始终对它们如何被用来控制、伤害或奴役保持警惕我们。 unstoppable, so it’s best to at minimum understand them, and perhaps even learn to exploit or live with them in certain circumstances, always remaining vigilant as to how they may be used to control, harm, or enslave us.
话虽这么说，在最早的阶段，它们总是显得“无害”，足以引起一些有限的兴奋，或者可能是对进展的好奇——毕竟，我们正在进入一个That being said, in the earliest stages they invariably appear ‘innocuous’ enough to entertain some limited excitement, or perhaps curiosity at the progression—after all, we 未知are 的前沿，这是我们许多人从小到大想象的一个令人兴奋的领域。我们青少年迷恋科幻小说等的痛苦。entering an unknown frontier, one which many of us grew up imagining, in the heady throes of our teenage infatuations with Science Fiction, and the like.
那么让我们继续讨论该技术的一些更吸引人的用途。So let’s move on to some of the more captivating usages of this technology.
首先，作为学习辅助工具，覆盖层已经可以用于学习工具等，如下所示：First, as learning aid, the overlays can already be used for such things as learning instruments, as follows:
这自然会发展成无穷无尽的应用程序，人工智能助手将能够实时“指导”你完成任何日常任务，例如步行/驾驶指导，以及外科医生执行复杂的手术。This will naturally blossom into endless applications where AI assistants will be able to ‘guide’ you in real time in doing any task as banal as walking/driving directions, to a surgeon performing complex operations.
主要的限制仍然是人们被迫佩戴的耳机笨重且不舒服。然而，随着每一代的迭代，它们不断变得更好、更小。最新的 Quest 3 已经比以前的型号小得多，侵入性更小，正如您在之前发布的视频之一中看到的那样，已经有智能眼镜开始进入市场，例如新的雷朋和 Meta 合作伙伴智能眼镜 – 然而这些设备目前还没有使用任何 AR/VR，而是使用了一些光记录和聊天机器人集成，这使得它们相当毫无价值。The chief limitation continues to be the bulky and uncomfortable headsets one is forced to wear. However, they continue to get better and smaller with each generational iteration. The latest Quest 3 are already much smaller and less invasive than previous models, and as you saw in one of the earlier posted videos, there are already Smart Glasses beginning to hit the market, like the new Ray Ban and Meta partnership Smart Glasses—however these don’t yet utilize any AR/VR but some light recording and Chatbot integration for now, making them rather worthless.
这个方向要困难得多，因为一副小眼镜缺乏重型处理能力，缺乏安装一副合适的 VR 护目镜所需的各种传感器和芯片的空间。然而，随着技术的进步，更多的东西可以容纳在更小的空间中，并且可以将所有处理工作转移到外部“云”单元中，该单元将信息传输回您的智能眼镜，从而使它们不再需要自己的笨重处理器。This direction is much harder because a small pair of glasses lacks the heavy-duty processing capabilities, the room for the various sensors and chips necessary to do what a pair of proper VR goggles can. However with technological advancement more can be fit into less space, as well as the possibility of offloading all the processing into an outside ‘cloud’ unit which streams the information back to your Smart Glasses, freeing them from the need of bulky processors of their own.
最终实现的一种方式是，“智能传感器”不仅会在我们的家里（如果你愿意的话）变得无处不在，更重要的是在现实世界中。街道、商店和公共场所可以配备传感器，例如可以从小眼镜读取队列，以便计算它们的方向，就像摄像机读取电影制作中的“跟踪球”一样：One way that will eventually take flight is likely that ‘smart sensors’ will become ubiquitous not only in our homes, if you so choose, but more importantly out in the real world. Streets, stores, and public places can have sensors which for instance can read queues from your small glasses in order to calculate their orientation in the same way that cameras read the ‘tracking balls’ in filmmaking:
这些无处不在的传感器可以完成 VR 耳机上笨重的内部硬件现在所做的所有方向处理，然后在您四处走动时将信号无线传输到您的“智能眼镜”。These ubiquitous sensors can then do all the orientation processing that the bulky internal hardware now does on VR headsets, then simply beam the signal wirelessly to your ‘Smart Glasses’ as you walk around.
当然，最终甚至不需要这样做，因为像马斯克即将推出的 Neuralink 这样的眼睛植入物和大脑植入物可能能够将覆盖层直接输入我们的视网膜和/或大脑。Of course eventually, there will be no need for even that, as eye implants and brain implants like Musk’s upcoming Neuralink will potentially be able to feed the overlays directly into our retinas and/or brains.
但截至目前，我个人会佩戴智能眼镜吗？不太确定，因为从附近的“云”源发射高能信号可能意味着大量射频/电磁辐射波直接进入您的大脑，这与长时间将手机靠近耳朵没有什么不同。但这也可以通过有线/电缆连接而不是无线传输来克服，至少对于家庭使用来说是这样。But as of now, would I personally wear a set of Smart Glasses? Not so sure, given that beaming energetic signals from a ‘cloud’ source nearby would likely mean the high emission of RF/EMR waves straight into your brain, no different than holding a cellphone close to ear for extended periods. But this can also be overcome with a wired/cable attachment rather than wireless transmission, at least for home use.
另外，作为进一步的警告。我经常提到 Quest 3，因为它恰好是这些最新技术的最新、最好的代表。我建议你去买一个吗？Also, as a further caveat. I’m bringing up the Quest 3 a lot because it happens to be the latest and best representative of these newest technologies. Do I recommend you go buy one?
….不….not 完全是。quite.
首先，Quest VR 是由 Meta 制作的，Meta 是世界上最卑鄙的公司之一，由一个几乎没有人性的全球主义技术官僚精神病患者经营。这就是意识形态的角度。First of all, Quest VR is made by Meta—one of the most despicable companies in the world, run by a barely human globalist technocratic psychopath of the worst persuasion. That’s the ideological angle.
实际的角度是，我上次听说，Quest 要求您拥有一个实际的、合法的“Facebook”帐户才能登录并使用他们的产品。我自己没有尝试过——我拥有不同品牌的 VR 耳机——也不使用 FB，但我听说它必须是一个与你的真实姓名相关联的可验证帐户，并且你无法被绕过或欺骗带有一些带有燃烧器地址的马甲仿冒品。也许有办法解决这个问题，但使用由扎克伯格的 Meta 制造或与之相关的任何产品或硬件无疑是一个巨大的障碍。The practical angle is that, last I heard, Quest requires you to have an actual, legit “Facebook” account in order to log into and use their product. I haven’t tried it myself—I own a different brand of VR headsets—nor do I use FB, but I’ve heard that it has to be a verifiable account linked to your real name and you can’t be bypassed or spoofed with some sockpuppet knock-off with a burner address. Maybe there’s ways to get around that, but it’s certainly a huge turn off from using any product or hardware made by, or associated with, Zuckerberg’s Meta.
因此，这并不是对任何特定产品的认可，而是对技术快速发展方向的探索。Thus, none of this is an endorsement of any particular product but rather an exploration of where the technology is rapidly heading.
我们在同一页上吗？好，我们继续吧！We on the same page? Good, let’s move on!
现在事情变得更加有趣，或者说……反乌托邦反常。Now things get even more interesting, or…dystopianly perverse.
苹果发布了 iPhone 15 Pro，它将Apple announced the iPhone 15 Pro, which will 为 Apple Vision Pro混合现实耳机捕捉3D“空间视频”，让用户capture以新的方式重温回忆。 3D ‘spatial videos’ for the Apple Vision Pro mixed-reality headset, allowing users to relive memories in a new way.
人们越来越多地推动能够捕捉你周围空间信息的技术，前面提到的 Quest 等已经可以做到这一点。但这可以配对并转化为一些“有趣”的方向。There are more and more pushes for technology that can capture the spatial information around you, which the aforementioned Quest amongst others can already do. But this can be paired and transformed into some ‘interesting’ directions.
例如，一位用户扫描了他的家并将其转换为数字建筑蓝图，该蓝图可以以多种方式使用，从将您自己的房子变成游戏环境到我们很快就会谈到的其他奇怪的事情。For instance, one user has scanned his home and transformed it into a digital architectural blueprint, which can be used in a variety of ways, from turning your own house into a gaming environment to other bizarre things we’ll get to shortly.
一些人认为，这些技术现在可以用来捕捉一个人的整个生活，或者至少是其部分生活的完整空间细节。这允许创建幽灵记忆，甚至可以通过 AR 将其填充到您的实际空间中。Some believe that these technologies can now be used to capture one’s entire life, or at least portions of it in fully spatial detail. This allows the creation of ghost memories, which can even be populated into your actual space via AR.
这只是当前应用程序的粗略演示。互动回忆：This is just a crude demonstration with current apps. Interactive memories:
我确信 3D 重建技术的杀手级用例是I’m convinced the killer use case for 3d reconstruction tech is 记忆memory捕捉我的父母今年早些时候退休了，我已经让他们的家永远永垂不朽 更多照片扫描是合法的 我们今天可以使用的最面向未来的媒介 扫描所有空间/地点/事物 capture my parents retired earlier this year and i have immortalized their home forever more photo scanning is legit the most future proof medium we have access to today scan all the spaces/places/things
当您的宠物去世后，您可以在捕捉它们的空间视频后继续在家中重温它们：When your pet passes away, you can continue reliving them through your home after capturing spatial videos of them:
你在颤抖吗？还被击退吗？或者你会和我一起沿着黑暗、不确定的道路继续这段旅程，走向技术奇点和变形吗？Are you shuddering? Repulsed yet? Or will you continue this journey down the dark, uncertain path toward technological singularity and transfiguration with me?
回想一下《银翼杀手 2049》中与虚拟 AI 女友 Joi 的场景Recall the scenes in Blade Runner 2049, with the virtual AI girlfriend, 。Joi.
或者《少数派报告》中汤姆·克鲁斯扮演的角色观看“全息记忆”的场景怎么样？Or how about the scenes in Minority Report where Tom Cruise’s character watches ‘hologram memories’.
但这些大多是普通的应用程序。But these are mostly mundane applications.
当您将新兴的人工智能功能应用于该技术时，该技术的阴暗面就会出现。The darker side of the tech comes when you apply burgeoning AI capabilities to it.
那些关注发展的人会知道，生成式人工智能已经越来越擅长模仿人类，无论是著名艺术家的风格，还是作者的写作商标。人类的相似性和声音也发生了同样的情况。Those who’ve followed the developments will know that generative AI has gotten increasingly proficient at mimicking humans, whether it’s a famous artist’s style or even an author’s writing trademarks. The same has been going on with human likenesses and voices.
人工智能现在可以合成任何人类声音，几乎完美地重新创建它，具体取决于您向其提供的目标人的语音数据量。AI can now synthesize any human voice, recreating it almost perfectly, depending how much voice data from the target person you feed it.
当你将这些东西结合起来时，你很快就能为你所爱的人创建不死的“化身”，他们的外观和声音都像你的爷爷，并且可以通过 AR 集成填充你的家 – 不仅仅是潜在的、静态的“记忆”，而是作为实际的交互式守护进程。When you combine these things, you will soon be able to create undead ‘avatars’ of your loved ones who look and sound just like your grandpaw, and can populate your home via AR integration—not just as latent, static ‘memories’, but as actual interactive daemons.
这已经在进行中。This is already underway.
上述《纽约邮报》文章指出：The above NYPost article states:
创立了多个人工智能平台的硅谷计算机科学家普拉蒂克·德赛博士大胆预测，到今年年底，人类的“意识”可能会被上传到数字设备上。Dr. Pratik Desai, a Silicon Valley computer scientist who has founded multiple Artificial Intelligence platforms, boldly predicts that a human being’s “consciousness” could be uploaded onto digital devices by the end of the year.
“开始定期记录你的父母、长辈和亲人，”“Start regularly recording your parents, elders and loved ones,”他周五在 Twitter 帖子中敦促道，该帖子已获得超过 570 万次浏览和数万条回复。 he urged Friday in a Twitter thread that’s since racked up more 5.7 million views and tens of thousands of responses.
“有了足够的文字记录数据、新的语音合成和视频模型，他们有 100% 的机会在离开肉身后永远和你在一起，”德赛继续说道。“到今年年底这应该是可能的。”“With enough transcript data, new voice synthesis and video models, there is a 100% chance that they will live with you forever after leaving physical body,” Desai continued. “This should be even possible by end of the year.”
像下面的视频这样的新技术已经允许人们立即重新创建 3D 对象，在 AR 空间中实时复制、操作和更改它们：New technologies like the following video, are already allowing people to instantly recreate 3D objects, duplicating, manipulating, and altering them in real time in AR space:
您很快就可以“扫描”一个人，并使用类似于 ChatGPT 的生成技术，重新创建他们的化身，以便在您选择的 AR 环境中无限期地生活，就像 BR2049 中的You’ll soon be able to “scan” a person and using generative ChatGPT-like technology, recreate an avatar of them to live indefinitely within your chosen AR environment, just like Joi一样Joi，只不过它可以是您所爱的人你的选择。 from BR2049—except it can be the loved one(s) of your choice.
一些公司Some companies 已经提供了already offer这种基本形式： basic forms of this:
我们大多数人都会这样做吗？可能不会——我确信它会在许多人（如果不是大多数人）中引起同样程度的恐惧和厌恶。但这很快就会发生。Would most of us ever do that? Probably not—I’m sure it elicits the same rank horror and repulsion amongst many if not most people. But it will be happening soon.
那么这些模仿声音和肖像的技术怎么样呢？这是最近的一个例子：So what about these technologies to mimic voices and likenesses? Here’s one recent example:
正如我们所说，已经开发出一些程序，可以将人工智能生成的嘴实时叠加到真实的人脸上，以改变说话者的声音和语言，同时真实地使他们自己的嘴匹配。这最初用于翻译目的，允许您以这样的方式翻译说话者，使其看起来像他们实际上在用所选语言说话：As we speak, there are already programs developed which can overlay AI generated mouths onto real human faces in real time in order to change the voice and language of the speaker while realistically conforming their own mouths to match. This is initially being used for translation purposes, allowing you to translate speakers in such a way that it looks like they’re actually speaking in the language of choice:
在上面的视频中，他警告说这项技术正在变得危险。他是对的，因为它In the above video, he warns that this technology is becoming dangerous. And he’s right in that it’s 已经already被用来在地缘政治领域达到不正当的目的。例如，就在上周乌克兰政治危机不断升级期间，一些此类深度造假视频令人们震惊，并有可能引发一系列令人不快的事件。 been used to deviously ends on the geopolitical spectrum. For instance, just last week during some of the escalating political crises in Ukraine, several such deepfake videos startled people and threatened to set off unpleasant chains of events.
其中一位是乌克兰将军扎卢日尼，他呼吁人们推翻泽连斯基：One was of Ukrainian general Zaluzhny who calls on people to go overthrow Zelensky:
事实上，在人们普遍愤怒之后，第二段视频甚至被发布，其中 DeepFaked Zaluzhny 的表现更加出色：In fact, after people widespread outrage, a second video was even released where the DeepFaked Zaluzhny doubles down:
问题是，人们必须明白，这不是传统的“DeepFake”，在传统的“DeepFake”中，你需要一个次要演员，然后煞费苦心地用算法替换他的脸。这些是 Zaluzhny 真实的、真实的、之前的视频，人工智能已经操纵嘴巴来匹配一组新的口语单词。这在很多方面都更加危险，因为它比传统的 DeepFake 更真实，因为人物本身的一切都是真实的，只有实时进行的嘴部微妙操作。The problem is, people have to understand that this isn’t a traditional ‘DeepFake’ where you get a secondary actor and painstakingly replace his face with an algorithm. These are actual, real, previous videos of Zaluzhny whereupon AI has manipulated the mouth to match a new set of spoken words. This is in many ways more dangerous as it’s far more realistic than traditional DeepFake, as everything about the personage himself is real, with only subtle manipulations of the mouth being done in real time.
这项技术成为全球热点和虚假信号的主要驱动因素只是时间问题——事实上可能已经在我们眼皮子底下做到了。It’s only a matter of time before this technology becomes the primary driver of global flashpoints and falseflags—and in fact may have already done, under our noses.
拜登政府似乎所做的为数不多的有用的事情之一就是One of the few useful things the Biden administration appears to have done is issuing a 两周前发布了一项新的行政命令，旨在减少人工智能即将带来的一些危险：new executive order just two weeks ago that seeks to curtail some of these coming dangers of AI:
除了据称为人工智能开发人员制定安全标准以防止人工智能控制生物武器开发等事情之外，它还通过建立某些“标准”来针对上述类型的 DeepFakes 采取行动，其中包括人工智能的身份验证和水印-生成的内容：Apart from supposedly setting safety standards for AI developers to, for instance, prevent things like AI getting control of biological weapons developments, it also takes action against the type of DeepFakes discussed above by establishing certain ‘standards’, which include authentication and watermarking of AI-generated content:
通过建立检测人工智能生成的内容和验证官方内容的标准和最佳实践，保护美国人免受人工智能驱动的欺诈和欺骗Protect Americans from AI-enabled fraud and deception by establishing standards and best practices for detecting AI-generated content and authenticating official content。商务部将制定内容认证和水印指南，以明确标记人工智能生成的内容。联邦机构将使用这些工具让美国人轻松知道他们从政府收到的通信是真实的，并为世界各地的私营部门和政府树立榜样。. The Department of Commerce will develop guidance for content authentication and watermarking to clearly label AI-generated content. Federal agencies will use these tools to make it easy for Americans to know that the communications they receive from their government are authentic—and set an example for the private sector and governments around the world.
不知道他们计划如何做到这一点，也不知道这是否只是为了获得更多控制权而做的遮羞布，就像政府通常的动机一样——我们必须拭目以待。No idea how they plan to do that and whether it’s just a figleaf to assume more control, as is the usual governmental motive—we’ll have to wait and see.
大多数不温不火地关注发展的人都知道，生成式人工智能今年发展非常迅速。我们从 Dalle 和 MidJourney 的早期版本（可以创造一些过时的新奇事物），到新的高度先进的多模式更新，可以进行微小的更改、识别和描述上传的照片等。Most who’ve tepidly followed developments know that generative AI has been advancing very rapidly this year. We went from early versions of Dalle and MidJourney which can create some passing novelties, to new highly advanced multimodal updates that can make minute changes, identify and describe uploaded photos, etc.
在这一点上，最新的聊天机器人具有“愿景”，这开辟了一系列全新的可能性。借助 Google 的“Bard”，您可以上传整个文档并让它对其进行总结或与您讨论细节。通过微软的“Bing”，您可以上传照片并让机器人向您描述甚至解释它们。On that note, the latest Chatbots have “vision”, which opens up a whole new range of possibilities. With Google’s ‘Bard’ you can upload entire documents and have it summarize them or discuss the finer points with you. With Microsoft’s ‘Bing’, you can upload photos and have the bot describe or even explain them to you.
例如，一种新的最受欢迎的消遣方式是上传模因并要求人工智能对其进行解释，尤其是那些带有某种图表或比较表的模因。For instance, one new favorite pastime has become uploading memes and asking the AI to explain them, particularly ones that have charts or comparative tables of some sort.
但这一领域的发展一直在加速。新系统和应用程序已经将“视觉”提升到实时水平，其中不仅包括照片，还包括视频。But the developments in this field have been accelerating. New systems and apps are already taking ‘vision’ to a realtime level, which includes not just photos but videos.
一个已经可用的 GPT 机器人示例，可以通过网络摄像头描述事物：One already available example of a GPT bot that can and describe things through your webcam:
另一种实验性人工智能已被用于通过合成人工智能语音对足球比赛提供实时解说：Another experimental AI has been used to provide realtime commentary on a soccer match with a synthetic AI voice:
新一代聊天机器人克服了以前的许多问题，例如臭名昭著的延迟，即它们需要几秒钟或更长时间才能对所提出的查询生成响应。And the new generation of Chatbots overcomes many of the previous problems, such as the notorious latency, whereby it takes several seconds or more for them to generate a response to an asked query.
在此示例中，人工智能电话营销员能够呼叫人类并与他们进行完全互动的对话：In this example, an AI telemarketer is able to call a human and have a fully interactive conversation with them:
当然，还有许多问题需要解决。但这些都在进行微调，并且新版本的示例已经比上述版本更先进。There are still many kinks to be worked out, of course. But these are all being finetuned, and there are samples out there of new versions already even more advanced than the above.
上周，Last week 一段广为流传的视频a viral video显示，一名男子在洛克希德·马丁公司的航空航天工程远程面试中表现出色，他对该主题没有任何实际了解，同时在第二台显示器上使用 ChatGPT 即时回答面试官提出的所有问题。许多人对这一先例感到愤怒，因为它可能会让人们通过撒谎进入危及很多人生命的工作——比如空中交通管制员等。 showed a man acing an aerospace engineering remote job interview at Lockheed Martin, with no actual knowledge on the topic whatsoever, all while using ChatGPT on a second monitor to instantaneously answer all the questions the interviewer was asking him. Many were outraged at how ‘dangerous’ a precedent this sets, as it could allow people to lie their way into jobs that have a lot of people’s lives on the line—like perhaps that of air traffic controller, etc.
借助新的“视觉”模式，这些人工智能机器人在识别疾病方面比以往任何时候都更加出色。你可以上传疣的照片并让它对你进行诊断，许多人发现它们比人类医生更准确，而人类医生曾经拥有腐败的偏见并受到各种类型的信息盲点的困扰。With the new ‘vision’ modes, these AI bots are getting better than ever at even identifying diseases. You can upload photos of your wart and have it diagnose you, and many people have found them to be more accurate than human doctors, ever possessed of their corruptible bias and plagued by informational blind spots of all types.
但最后的前沿将是将新的、最先进的 GPT 与实际的物理机器人化身相结合。先进的推理能力现在可以通知物理机器采取行动并创造出期待已久的、接近意识的机器人。But the final frontier will be combining the new, most advanced GPTs with actual physical robotic avatars. The advanced reasoning capabilities can now inform the physical machine to act and create the long-awaited, approaching-consciousness robots.
其中一个例子One example of this 是 ChatGPT 变体与臭名昭著的波士顿动力式机器狗之一的配对，但这只是一个快速而肮脏的测试，相对不那么令人印象深刻。was a pairing of a ChatGPT variant with one of the infamous Boston Dynamics-style robot dogs, but that was just a quick and dirty test, relatively unimpressive.
更令人印象深刻的是即将推出的专业消费级产品，例如 OpenAI 的“Eve”，据说它将其旗舰产品 ChatGPT 与实际的自主机器人结合起来，这里展示了其触觉能力：More impressive will be the upcoming professional consumer level products like OpenAI’s ‘Eve’ which is said to combine their flagship ChatGPT with an actual autonomous robot, seen here demonstrating its tactile abilities:
众所周知，波士顿动力公司在高度敏捷的机器人领域处于领先地位，例如他们的 Atlas 系列：As many know, Boston Dynamics is leading the field in highly agile robots like their Atlas series:
现在想象一下将上述功能与下一代近意识人工智能功能配对。在监管允许的情况下，在短短几年内，它们可能会改变我们周围的世界。一些人认为它们有一些关键的潜在缺点，例如电池寿命短，这些缺点将阻止此类机器的普及。但随着这些技术变得无处不在，这个问题在未来很容易得到解决。想象一下一个高科技城市，周围遍布着相当于“充电站”的东西。除了不“充电”之外，他们只会分发“可租赁”电池组，这些电池组将立即与数以百万计的此类机器人/机器在运行中切换。每个角落都可以有它们，如果你的机器人助手电量“低”，他会很快在街角分配器中用新电池组更换电池组，然后继续前进，同时可能会自动向你的帐户收取一些象征性的费用收费。Now imagine pairing the above with the next generation of near-conscious AI capabilities. In only a few short years they will likely transform the world around us, regulation permitting. Some believe they have a few key underlying shortcomings which will prevent such machines from ever becoming widespread, like for instance short battery life. But this will easily be solved in the future as these technologies become ubiquitous. Imagine a high-tech city that has the equivalent of ‘charging stations’ placed all around. Except instead of ‘charging’, they will merely dispense ‘rentable’ battery packs that will be instantly switched out with millions of such robots/machines on the fly. Every corner can have them and if your robot sidekick is ‘low’ on juice, he’ll just quickly exchange his battery pack with a fresh one in a street corner dispenser and continue on his way, while perhaps automatically incurring your account for some nominal charging fee.
但最大的新威胁是左翼审查制度、意识形态和社会文化工程的逐步压制和硬连接到下一代人工智能产品中。But the biggest new threat is the progressive clampdown and hardwiring of Leftist censorship, ideology, and socio-cultural engineering into the next generation of AI products. 克里斯托弗·鲁福解释了Christopher Rufo explained拜登的新举措： Biden’s new initiative:
拜登总统签署了一项行政命令，要求President Biden has signed an executive order that will 人工智能公司“解决算法歧视”并“确保人工智能促进公平”。require AI companies to “address algorithmic discrimination” and “ensure that AI advances equity.”他们希望将 CRT 和 DEI 的原理嵌入到人工智能的各个方面。 They want to embed the principles of CRT and DEI into every aspect of AI.
你明白了吗？将硬件 DEI（多样性、公平性和包容性）和 CRT 原则融入人工智能，使其更加“包容”。特别注意关于解决“算法歧视”的那句话，这基本上意味着对人工智能进行编程来模仿目前用来窒息西方世界的专制大厅监控管理主义。Did you catch that? Hardwire DEI (Diversity, Equity, and Inclusion) and CRT principles into AI to make it more, well, “inclusive.” Particularly note the line about addressing “algorithmic discrimination” which basically means programming AI to mimic the present tyrannical hall-monitor managerialism being used to suffocate the Western world.
对于 GPT 程序的狂热用户来说，您会注意到这已经成为一个问题，因为聊天机器人在推动某些叙述方面变得非常顽强，并确保您不会对历史事件的任何不方便的解释做出错误的思考。For avid users of GPT programs, you’ll note this is already becoming a problem, as the Chatbots get extremely tenacious in pushing certain narratives and making sure you don’t commit WrongThink on any inconvenient interpretations of historical events.
埃隆·马斯克（Elon Musk）所谓的反击新尝试已经产生了他自己的聊天机器人技术变体，名为Elon Musk’s new attempt to supposedly push back has yielded his own variation on Chatbot tech named Grok Grok，据说这是一个不敬的“基础”和毫无歉意的政治不正确模型。但我还没有机会亲自测试它，看看它如何应对真正棘手或有争议的场景。, which is said to be an irreverently ‘based’ and unapologetically politically incorrect model. But I haven’t had the chance to test it out myself yet to see how it responds to truly tricky or controversial scenarios.
无论如何，这基本上并不重要，因为任何此类异端模式都可能仍然是新鲜事物，而真正的大垄断企业和巨型企业将与被俘虏的政府合作，以确保It mostly doesn’t matter anyway because any such heterodox models will likely remain a novelty while the truly big monopolies and megacorps will work with captured governments to make certain that 他们的their压迫性模式将成为我们社会中无处不在的黄金标准，没有任何限制。第二个选项。这些将融入我们的生活，模糊公共“公用事业”和私营公司之间的界限，正如社交媒体巨头已经熟练地做到的那样。 oppressive models will be the ubiquitous gold standard to be mandated in our societies with no second options. These will become embedded into our lives, blurring the lines between public ‘utility’ and private corporation as social media giants have already so expertly done.
需要澄清的是，就像医学或广播一样，将会有类似于 FCC、FDA 等的政府监管机构，它们将监管允许广泛使用的“标准”代码框架和算法。这些框架将由大型科技巨头运营，并将有强制性的“道德/公平/包容”代码库编程到其中。这些“标准”人工智能将是唯一允许扩散的人工智能，并成为事实上的垄断者，将运行我们所有的“物联网”、智能家居等，就像微软的Windows等一样。To clarify, just like with medicine or broadcasting, there will be government regulators akin to the FCC, FDA, etc., which will regulate the ‘standard’ code frameworks and algorithms allowed to be used on a wide scale basis. Those frameworks will be the ones operated by the big tech giants, and will have mandatory ‘ethic/equity/inclusion’ codebases programmed into them. These ‘standard’ AIs will be the only ones allowed to proliferate and become the de facto monopolies that will run all of our “internet of things,” smart homes, etc., just like Microsoft’s Windows, etc.
这样的人工智能将确保始终巧妙地引导公众话语，即允许的奥弗顿窗口，以便叙事流程永远不会超出舒适的范围。AR/MR/ER 的发展将协同工作，通过删除或替换现实中难看或不方便的部分，将它们从视线中抹去，从而创造实时“公平”。Such AIs will make sure to always subtly steer public discourse, the allowable Overton windows, so that the narrative flow never exceeds comfortable bounds. The AR/MR/ER developments will work in tandem toward creating realtime ‘equity’ by deleting or replacing unsightly or inconvenient parts of reality on the fly, wiping them from sight.
Of course, that’s the near future envisioned by 当然，这是他们them所设想的不久的将来，但它仍然有可能失败，至少在“西方”集体中除了最妥协的国家之外的所有国家都是如此。—but there’s still a chance it will fail, at least in all but the most compromised countries of the collective ‘West.’
现在，趁你还有机会，欣赏我们未来霸主的不打扰前辈吧：For now, enjoy the unintrusive predecessors of our future overlords while you still have the chance:
