A closer look at Anthropic's $100 billion public-good gamble

So, what makes Anthropic's plan different? First, Anthropic says it has created a completely separate organization to serve the public good. The company plans to give this organization 5% ownership of the company in shares, subject to approval by its board. The new organization is set up as a limited liability company, or LLC, rather than a public benefit corporation. Anthropic says this structure gives it more freedom in how it can help people.

This structure would let Anthropic give help in many different ways. The company also says there is no political agenda behind the fund. Its only goal is to benefit the public. But exactly how it will do that is perhaps the most interesting and least clear part of the announcement. Anthropic says it may make large donations to existing nonprofit groups or give money directly to certain groups of people.

You could argue that Anthropic has some of the best understanding of AI because it is building these models itself. The company has carried out several research studies that ask people who use Claude, as well as people who have not used it, how AI might affect them. To Anthropic's credit, it has been very active and open about sharing the results of this research. This means the company may have a good understanding of how AI models could affect people around the world.

Still, Anthropic cannot know everything about what will happen. The only way to really understand AI's effects is to release these models to people and watch what happens over the next few years. Jack Clark, one of Anthropic's co-founders, has said that the fund is a deliberate effort to share some of the money and wealth created by Anthropic's models with humanity.

The fund is also meant to act like insurance against the problems and major changes AI could cause. AI models might replace many jobs, take away people's sense of purpose, or make it harder for people to earn a living. It is important to understand that this is not universal basic income, which means giving people regular payments. Instead, the plan is to give away 5% of the company's ownership now, with a promise to give more if the company's value grows. For every 50% increase in its value,

Anthropic will add another 2% of the company's ownership. One thing I found interesting is where these shares will come from. The first 5% will be given as a donation, and so will each extra 2% whenever the company's value rises by another 50%. What I had not realized is that Anthropic has committed up to 20% of the company's ownership to the public good. That is an enormous amount, especially for a company described as the fastest-growing private company ever.

The story behind these shares is even more interesting. When Anthropic's founders started the company, they expected that powerful AI might create serious problems or challenges that they would need to prepare for. This decision was made before Claude even existed. The founders committed 80% of their own shares to helping the wider public.

They also created a matching program in which employees can choose to give some of their company shares to support the public good. About 50% of Anthropic's employees take part in this program. Although Anthropic has not existed for very long, people connected with the company have committed 20% of its ownership to public benefit. The company has also promised that it may give another 2% for every 50% increase in its value.

This is a kind of generosity that we have not seen from other companies before. It is important to understand what these numbers could mean in money. I did some calculations, and if Anthropic reaches a value of $2 trillion, which is the rumored value for its planned initial public offering, or IPO, then 5% of the company would be worth $100 billion. But