openai的宫斗戏

题图：cheva 用必应DALL E3创作

这个世界除了一些打打杀杀和政治上的尔虞我诈之外，也还是有一些精彩的事情。比如最近几天就有两件出乎人意料的事情发生了。

一件是具有奥派思想的经济学家米莱刚当选阿根廷的总统。他的政治主张中有非常多的符合奥地利学派经济学的主张。比如他要大幅度的精简政府，把20个政府部门拆到8个，还要废除中央银行。这无疑是最大胆的，也是最贴近奥派思想核心的政治主张。

奥派就是主张完全的自由市场经济，包括货币也应该是由自由市场自行产生，而不是由国家和政府权威或者政治权威来指定。事实上人类历史上使用的绝大部分货币都是由市场自发选择产生的。从金银到最早的纸币、铜钱都是在没有政府干预下的自由市场自发形成的。

另一件是AI业界的领头羊，如日中天的公司OpenAI，GPT-4的创立者，居然一夜之间分崩离析了，情节离奇犹如风云大戏。因为就在不久前，OpenAI才刚刚对它们的GPT-4进行了一次大升级，推出了可以音对话和读图的多模态GPT-4和可以发布自己的人工智能机器人应用的GPT-S，还有速度更快、价格更低的GPT-4 Turbo API，怎么看都是蒸蒸日上的，而且还甩开了它的竞争者一大段的距离。可以说在这场AI大潮中，只有和OpenAI保持合作关系的微软跟上了它，而谷歌已经被大大的甩在了身后。

而从公开媒体信息来看，OpenAI成功的背后的灵魂人物是一位年轻人，名字和打怪兽的奥特曼非常相似，也就是萨姆 奥特曼，他是OpenAI的CEO。当然也有人打趣地说CEO名字非常好听，但实际上也就是高级打工仔，随时可以被老板开掉，没想到这样的明星级的人物，可以说是公司灵魂的人也会碰到这样的情况。

就是前天，奥特曼突然接到通知要他参加一个在用Google Meeting召开的线上会议，而就在这个会上，他才被告知董事会已经做出了决定将他开除。可想而知，和所有局外人一样，萨姆 奥特曼本人也是一头雾水。

目前为什么会发生这场事变，原因众说纷纭，有的说是因为董事会的高层和奥特曼发生了理念的冲突，不希望他再训练比GPT-4能力更强的类似GPT-5这样的产品，认为会给人类带来灾难。当然，也许真实的原因并没有这么复杂，也许仅仅是人性在作怪而已。

开除公司的重要人物似乎不是一个稳健经营的公司所能够做出的，而是像是公司中某个大权在握的人草率地做出的即兴的决定。就像川普在任总统的时期，也曾用类似的方式开除了中央情报局的局长科米，而科米本人也是在电视新闻上得知总统的这一决定的。

说到底，OpenAI一开始是一个非营利的机构，它所秉持的理念是，像AI这样的技术，不应该被小部分有钱的大公司所掌握，它应该属于每一个人。所以它一开始的研究成果也是完全开源的，比如GPT-4的老前辈GPT-3，就是一个开源的语言模型。还有，现在我也经常在使用的Whisper语音转文字模型，也是在开源的语音转文字的模型中效果最好的一个，比网上那些需要花钱订阅才能使用的语音转文字模型要好用得多，而且还免费。

但是随着GPT-4大火之后，公司逐渐放弃了开源的理念，转向了商业模式，但是组织架构并没有发生多大改变，还是沿用原来的非营利组织的架构。由三个人组成的董事会拥有非常大的权力，而股东几乎没有任何的权力，股东和投资者则没有任何话语权，所以整个事件的背后应该就是一场宫斗大戏。OpenAI的高层看到了组织发明的产品具有巨大的商业和经济利益，于是利用手中的权力想多占，于是赶走了对自己威胁最大的人。

我们这些受过九年义务教育的中国人都知道这样一句话，就是经济基础决定上层建筑，这话到底对不对，其实也还是有待商榷的，但是有一点是毋庸置疑的，经济基础和上层建筑必须是相适应的，他们到底谁决定谁不知道，但至少有一点必须相互适应。

对OpenAI的事情来说，要么你就坚持自己的非盈利组织的理念，不要将GPT-4商业化，要么你就改变组织架构，采取商业公司的股份公司的架构来运作，以非盈利组织的架构来运营商业公司，或者以商业公司的架构来运营非盈利的事业，这两者都将会导致灾难。

这两件事件，无论是奥派总统的出现，还是OpenAI的垮台，都给我们带来了很多的思考。我们应该如何看待自由市场经济的优势和风险？我们应该如何看待AI技术的发展和影响？我们应该如何看待组织架构和理念的匹配和变化？这些问题都没有简单的答案，但是我们不能回避它们，因为它们关乎着我们的未来。

---

![OIG (7).jpeg] (https://cdn.steemitimages.com/DQmXEttFLdVowjeiJsRNSUuAMsf1jaEjvdZijzZeeKcU2Da/OIG%20(7).jpeg)

Title image: cheva created with Bing DALL E3

In addition to some fights and killings and political deceit, there are still some wonderful things in this world. For example, in the last few days, two unexpected things have happened.

One is that Milley, an Austro-minded economist, has just been elected president of Argentina. Many of his political ideas are in line with the ideas of the Austrian School of Economics. For example, he wants to drastically streamline the government, dismantling 20 government departments into 8, and abolishing the central bank. This is undoubtedly the most daring political proposition, and it is also the political proposition closest to the core of Austrian thought.

The Austrians advocated a complete free market economy, including that money should be generated by the free market itself, rather than designated by state and government authorities or political authorities. In fact, the vast majority of currencies used throughout human history have been produced by spontaneous choice by the market. From gold and silver to the earliest paper money and copper coins, they were all formed spontaneously in the free market without government intervention.

The other is that the leader of the AI industry, such as OpenAI, the founder of GPT-4, fell apart overnight, and the plot was bizarre. Because not long ago, OpenAI just made a big upgrade to their GPT-4, launching a multi-modal GPT-4 that can read audio and read pictures, and GPT-S that can release its own artificial intelligence robot applications, as well as the faster and cheaper GPT-4 Turbo API, no matter how you look at it, it is thriving, and it is also a long distance away from its competitors. It can be said that in this AI tide, only Microsoft, which maintains a cooperative relationship with OpenAI, has kept up with it, and Google has been greatly left behind.

Judging from public media information, the soul behind OpenAI's success is a young man whose name is very similar to Ultraman who fights monsters, that is, Sam Ultraman, who is the CEO of OpenAI. Of course, some people joke that the CEO's name is very good, but in fact, he is a senior worker who can be fired by the boss at any time.

It was the day before yesterday that Ultraman was suddenly notified that he was attending an online meeting using Google Meeting, and it was at this meeting that he was informed that the board of directors had made a decision to expel him. As you can imagine, like all outsiders, Ultraman Sam himself is confused.

There are different theories as to why this incident is happening, but some say that it is because the top of the board of directors has a conflict of ideas with Ultraman, and they don't want him to train a product like GPT-5 that is more powerful than GPT-4, believing that it will bring disaster to mankind. Of course, maybe the real reason is not so complicated, maybe it's just human nature at work.

Firing an important person in a company does not seem to be something that a well-run company can make, but rather an impromptu decision made by someone in a powerful position in the company. Just like when Trump was president, he fired CIA Director Comey in a similar way, and Comey himself learned of the president's decision on the TV news.

At the end of the day, OpenAI started out as a non-profit organization with the idea that technology like AI shouldn't be in the hands of a small group of big companies with money, it should belong to everyone. Therefore, its initial research results are also completely open-source, such as GPT-4's predecessor GPT-3, which is an open-source language model. Also, the Whisper speech-to-text model that I use a lot now, and it's also the best of the open-source speech-to-text models, which is much better than the speech-to-text models that require a subscription on the Internet to use, and it's free.

However, after the fire of GPT-4, the company gradually abandoned the concept of open source and turned to a business model, but the organizational structure did not change much, and the original non-profit organization structure was still used. The board of directors composed of three people has very much power, while the shareholders have almost no power, and the shareholders and investors have no say, so behind the whole incident should be a palace fight. OpenAI's top management saw that the products invented by the organization had huge commercial and economic benefits, so they used the power in their hands to take more of it, so they drove out the people who posed the greatest threat to them.

Those of us who have received nine years of compulsory education Chinese all know that the economic base determines the superstructure, whether this is true or not is still up for debate, but there is no doubt that the economic base and the superstructure must be compatible, and they do not know who decides who decides, but at least one thing must be adapted to each other.

In the case of OpenAI, either you stick to your non-profit philosophy and don't commercialize GPT-4, or you change the organizational structure and operate as a joint-stock company of a commercial company, run a commercial company as a non-profit organization, or run a non-profit business with the structure of a commercial company, both of which will lead to disaster.

These two events, whether it is the emergence of the Austrian president or the fall of OpenAI, have brought us a lot to think about. How should we view the advantages and risks of a free market economy? How should we view the development and impact of AI technology? How should we view the alignment and change of organizational structure and philosophy? There are no easy answers to these questions, but we can't avoid them because they are related to our future.