You have a choice of having society be held accountable by an open source “machine” that anyone can read, deconstruct, scrutinize, test, debug and correct within a relatively short period of time or human beings.

您可以选择让开源“机器”对社会负责，任何人都可以在相对短的时间内或人类中阅读，解构，检查，测试，调试和纠正开源 。

Human minds are, for the foreseeable future, inscrutable neural networks comprised of a never-ending, psycho-social labyrinth of biases, stereotypes and prejudices; manipulatable by bribes and blackmail; compromisable by everything under the sun, from conflicting economic interests to unresolved pyschosomatic childhood trauma to just being in a bad mood that day.

在可预见的未来，人的思想是难以理解的神经网络，由无休止的，偏见，刻板印象和偏见的社会心理迷宫组成。 可以通过贿赂和勒索操纵； 从冲突的经济利益到无法解决的童年性精神创伤，甚至那天心情不好，阳光下的一切都会损害它。

The System Must Be…1. Open Source and Open Data. All inputs and all outputs.2. All processes should be deterministic i.e. reproducible3. Output a cryptographic hash of its internal state for every decision4. Publicly Testable During Runtime

系统必须是…1。 开源和开放数据。 所有输入和所有输出2。 所有过程都应是确定性的，即可重复的3。 为每个决策输出其内部状态的加密哈希值4。 运行时可公开测试

One’s source code is permission based, the other is permissionless.

一个人的源代码是基于权限的，另一个是未经许可的。

One conceals their lives, the other is a completely autonomous and economically self-interested AI designed to give machines an intuitive understanding of the real world. And the byproduct of this artificial economic behaviour causes it to not only reveal everyone’s life to itself but freely to any other entity who observes its internal state. No more secrets.

一个掩盖了他们的生活，另一个掩盖了完全自治且经济上自私的AI，旨在使机器对现实世界有直观的了解。 这种人为的经济行为的副产品使它不仅向每个人展示自己的生活，而且向观察其内部状态的任何其他实体自由展示 。 没有更多的秘密。

Total accountability. But not necessarily to a human, to an open source “machine” that can see everyone.

全面问责制。 但不一定是人类，而是可以看到所有人的开源“机器”。

And at least with an AI, even if it’s been trained on a data set that makes it inherently biased against a particular demographic, it will at least reveal its biases to you. Moreover, if it’s been trained on human data, it may provide evidence that there are significant biases in your human decision making systems.

至少对于AI而言，即使对其进行了数据集训练，使其固有地偏向特定的人口统计特征，它也至少会向您显示偏见。 此外，如果经过人数据培训，则可能提供证据表明您的人为决策系统存在重大偏差。

But human algorithms are a completely different story. They know when they’re being tested. And thanks to a cruel, evolutionary conflict of interest, their reward function incentivizes them to conceal their most dangerous biases whenever they’re tested. Just so they can be put into production! And humans are bad judges of their own biases at the best of times. So if the testing algorithm is also human, you’ve got “the blind leading the blind”.

但是人工算法是一个完全不同的故事。 他们知道何时进行测试。 并且由于残酷的，进化的利益冲突， 他们的奖励功能激励他们在测试时隐瞒最危险的偏见 。 以便它们可以投入生产！ 人类在最佳时机是对自己偏见的错误判断。 因此，如果测试算法也很人性化，那么您就是“盲人领先盲人”。

你不知道的魔鬼 (The Devil You Don’t Know)

Many of us live in horror movies. Our parents lived in horror movies and their parents lived in horror movies… There are a few tropes you always notice in the horror genre. Elements that seem to speak volumes to the human psyche.

我们中的许多人都生活在恐怖电影中。 我们的父母生活在恐怖电影中，而他们的父母生活在恐怖电影中……在恐怖类型中，您总是会注意到一些比喻。 似乎对人类心理说得通的元素。

1. Abnormal Powers: To be a true horror film the monster must be “supernatural” in some form or another. That is to say, its power is errie and seemingly unstoppable by conventional means. A series of crises soon reveals to the story’s protagonists that the monster not only exists outside of the moral and political “social contract” they understand but that there aren’t any known weapons, let alone legal systems, that could possibly save them from it. Forming the central conflict of the genre.

   异常力量：要成为一部真实的恐怖片，怪物必须以某种形式“超自然”。 就是说，它的力量是错误的，并且似乎无法通过传统方式阻止。 一系列危机很快向故事的主人公揭示了这个怪物不仅存在于他们理解的道德和政治“社会契约”之外，而且还没有任何已知的武器，更不用说法律制度了，可以将他们从中拯救出来。 形成体裁的中心冲突。

2. Astonishing Incredulity: To make matters worse, every character for whom the monster is not personally injurious has to deny its existence. To the point that even if it’s blatantly obvious that a monster is killing people they have to deny what they see. They’ll consistently devise some counterfeit explanation so as not to discredit the social contract they believe is the universal principle governing their reality.

   令人难以置信的令人难以置信：更糟糕的是，每个怪物个人没有伤害的角色都必须否认其存在。 即使很明显地，一个怪物正在杀死人们，他们也不得不否认他们所看到的东西。 他们将一贯地设计一些伪造的解释，以免破坏他们认为是支配现实的普遍原则的社会契约。

3. Pursuit: The monster is never sedentary. It hunts the protagonists; they can never truly rest. Creating the “rising action” of the story.

   追求：怪物永远不会久坐。 它猎杀主角； 他们永远无法真正休息。 创造故事的“上升动作”。

4. An Abnormal Solution: Due to the dilemma caused by 1, 2 and 3, any protagonists that survive to the end will, by necessity, have come to the realization that the monster has to be addressed on its own terms. It can only be defeated using something that is the abnormal and supernatural complement to the horror’s own wretchedness. And as far as the monstrosity is concerned, it is even more unrelenting.

   异常解决方案：由于1、2和3造成的困境，任何生存到最后的主人公都必须意识到必须按自己的方式解决怪物。 只能使用恐怖本身的不正常和超自然的补充来击败它。 而且就怪兽而言，它更加坚定。