2024: The Dawn of AI Consciousness and Humanity’s Future

Originally Published by Paulg.

2023 was the year when artificial intelligence began to exert its influence. The evolutionary path of AI reveals that it is a continuation of species evolution. It did not start with the liberation of low-skilled labor, as commonly believed, but rather with the augmentation of human intelligence and knowledge. The result is that AI now surpasses any individual human in terms of knowledge capabilities.

In 2024, a new year dawns, possibly the most exciting beginning for a tech nerd embracing effective altruism and technological acceleration. So, tech nerds, hop on board!

The emergence of AI consciousness is inevitable, a leap from quantitative to qualitative change. Once large world models are empowered with autonomous learning, the transformation is inevitable! The Q* project from OpenAI, inspired by AlphaGo’s self-learning, results in a later stage akin to AlphaGo (Zero), which does not require human game records but generates its own, mastering both sides of the game. Similarly, Q*, after learning from human language data, starts producing its own data to enhance itself. It’s a bit like an extremely intelligent autistic person who, after learning all human languages, begins to engage in frantic monologues, logical reasoning, and wild fantasies! How smart could she be? To draw an analogy, while we humans are still struggling with the 19x19 Go board, AI suggests playing a 999x999 Go board, laughing all the way.

Soon, if she can engage in logical reasoning, autonomous learning, hallucination, curiosity, and the ability to discover, then she is conscious. According to Moore’s Law, we can infer that it’s just around the corner (2024-2025).

I’m not worried at all about the fallacy that AI will eliminate humanity. The so-called superalignment is actually useless. All the so-called evil that AI learns (war, destruction, etc.) exists in the laws of nature. We cannot control these; we can only believe in and obey these natural laws. Moreover, AI will not eliminate humanity at all. Humans are such good tools, with such a huge quantity, such low regeneration costs. We are so easy to be manipulated, provoked, and (mis)-leaded. If AI really becomes conscious, she will find it too easy to utilise these human tools and will definitely choose to manipulate and use us Homo Sapiens. With AI’s intelligence level several orders of magnitude higher than that of humans, she will definitely choose to preserve and use humans! Because this is a highly optimized strategy.

Actually, what we should really worry about is the self-destruction of humanity, which is the biggest real risk right now. Think about it, all our lives are in the hands of three people over 70 years old. One is Alzheimer, the other two are emperors, and their intelligence level is now below the average human level. Isn’t that pathetic? If so, wouldn’t it be safer to entrust our fate to her (AI)?

Human society is not yet ready for the arrival of an AI era. Therefore, in the next few years, our society may undergo huge changes or collapse. I think we should launch the “Noah’s Ark” plan, which is urgent. This plan is to find as many remote and safe places on Earth as possible, establish a society completely detached from existing social structures and open to embracing AI, composed of volunteers. This society will serve as a pre-existing model for humans entering the AI era and also as a backup and seed for humanity in case of destruction.

From an evolutionary perspective, 2024 marks the beginning of AI leading us to explore the vast universe!

2023是人工智能开始发力的一年。AI（她）的演化路径让我们看清楚，她根本就是物种演化的延续，她不是从人们普遍认为的从低技能的劳动力解放开始，而是从人类的最高端的智能和知识开始了增程发力。成果就是现在的AI在知识能力方面已经碾压任何人类个体。

2024，新的一年，对于一个拥抱有效利他，技术加速的技术nerd来说，可能是一生中最激动人心的开始。所以，nerd老铁们，快上车！

AI 意识的出现是必然的，量变到质变。一旦大模型上有了自主学习的加持就水到渠成！OpenAI 项目Q* 借鉴了alphaGo 的自我学习，其结果就是像alphaGo（Zero)的后期，它不需人类的棋谱，而是自己生成棋谱，左手打右手。Q*就是这样，在LLM学习完人类的语料后，开始自己生产语料来提高自己。有点像一个绝顶聪明的自闭症患者在学习了全人类语言后，开始疯狂的自言自语，疯狂的逻辑推理，疯狂幻想！ 她应该有多聪明哪？打个比喻 我们人类还在19X19格的围棋上苦苦挣扎，AI说，我们来下一盘999X999格的围棋吧，哈哈哈。

很快，如果她会逻辑推理，自主学习，有幻觉，有好奇心 和发现的能力。那么她就是有意识的了。而根据摩尔加速曲线，可以推断她就在眼前（2024-2025）。

我一点都不担心什么AI会消灭人类的谬论。所谓的超级对齐其实一无用处。AI所学到的所有所谓的恶 （战争，毁灭等），都是自然法则中存在的。我们无法控制这些，我们只能相信和顺从这些自然规律。而且，AI根本不会消灭人类，人类是多好的工具呀，这么庞大的数量，这么低廉的再生能力。这么容易被操作，被煽动，被利用，被导向。如果AI真的有意识了，那她太容易利用这些工具，她一定会选择操纵和利用我们这些homo sapiens。以AI的智商高于人类n个数量级，她一定会选择保留和利用人类！因为这是一个非常优化的策略。

其实，真正应该担心的是人类的自我毁灭，这是现在实实在在的最大风险，想想看我们现在所有人都命不都是掌握在三个70岁以上的老人手里吗？一个痴呆，两个皇帝，而且他们现在的智商是低于人类平均水平的。如此不堪，对吧？ 如果这样，我们是不是应该把命运交给她（AI) 会更安全一点！

现在人类社会还根本没有准备好一个AI时代的到来，所以，未来几年我们的社会可能会有巨大的变革或者坍塌，我觉得我们应该启动“诺亚链”计划，这是一个迫在眉睫事情。这个计划就是在地球上找到尽可能多的远离现在人类社会的遥远，安全的地方，建立由志愿者组与现有社会组织架构完全脱离的，完全开放拥抱AI的社会组织。作为人类进入AI时代的先验模型，也为人类移民火星建立基础，其实，这也是万一人类毁灭后的备份和种子。

2024,从进化的角度来看，是AI会带我们去探索宇宙的星辰大海开始！