Now that Sam Altman is back at OpenAI as the CEO, the rumours about the OpenAI-Anthropic deal is officially off the table. The current board consists of Bret Taylor, Larry Summers, and Adam D’Angelo.

According to earlier reports, the previous OpenAI board consisting of Helen Toner, Tasha McCauley, Ilya Sutskever, and also Adam D’Angelo approached Dario Amodei, the co-founder of Anthropic, for a potential merger. The idea was to replace Altman as the CEO with Amodei. Though it is not clear if there was any serious discussion about the deal, Amodei turned down the offer as he is already the CEO at Anthropic, which is a direct competitor to OpenAI.

The OG board loved Anthropic, OpenAI not so much 

Looking back though, the board that fired Altman from OpenAI had negative views about the technology that OpenAI is building. For example, Toner wrote a paper that was critical of OpenAI, and praised Anthropic. Toner seems to be one of the reasons that Altman was ousted from OpenAI.

Her paper said that Anthropic’s willingness to not release its AI models is “exactly the kind of frantic corner-cutting that the release of ChatGPT appeared to spur”. She further praised Anthropic for investing more into AI safety than OpenAI. People on X criticise her views, saying it sounds more like an opinion piece than a research paper.

Altman has complained that Toner’s paper was criticising OpenAI’s effort for keeping its AI safe, according to an email he wrote to his colleagues. He said that the paper was dangerous for the company’s reputation since a board member’s criticism “carries a lot of weight,” he said. Altman was definitely on the right over here. 

It all points to the fact that old OpenAI’s board was not aligned with the for-profit approach that Altman had steered the company towards. Or maybe, it was just a case of revenge by Toner for the criticism. 

Regardless, Anthropic partnering with OpenAI without Altman, would have been an ideal deal for the startup. Anthropic could have attracted the best AI talent in the world, while OpenAI board would be happy with the AI safety approach. 

Anthropic predicts it would generate a revenue stream of $200 million by the end of this year, translating to nearly $17 million monthly revenue. By the end of 2024, it hopes to generate $40 million in monthly revenue.

The only point of contention would be OpenAI’s non-profit approach and Anthropic’s for-profit one, which might also eventually change given that OpenAI’s board has resigned and Altman is back as the CEO. The deal is no longer happening. 

Why the deal was a good idea

Anthropic’s Claude, its rival to GPT-4, just got a massive upgrade. Claude 2.1, the latest iteration of its AI language model, is now available through API, revolutionising the claude.ai chat experience. Bringing forth key enhancements for enterprises, Claude 2.1 introduces a remarkable 200K token context window, substantial reductions in model hallucination rates, and a beta feature called tool use.

GPT-4 Turbo, announced at OpenAI DevDay earlier this month, boasted a 128k context length, which was a step up from Claude-2 holding a 100k context length. Now, the 200k context length announcement is clearly an indication that Anthropic is trying to sway away a lot of OpenAI customers amidst the chaos.

The deal between OpenAI and Anthropic would have been good for them competitively against Microsoft and Google. About 95% of the employees of OpenAI, who are considering quitting, are juggling between going to Microsoft or Anthropic, Cohere, Inflection. 

Dario and Daniela Amodei, the sibling co-founders of Anthropic, are former OpenAI employees. Daniela was the VP of safety and policy at OpenAI and Dario was the vice president of research. Both of them have been quite outspoken about the need for caution in AI development and increasing regulations to make AI systems safe.

This is possibly one of the reasons why the duo departed from OpenAI to form Anthropic. The two-year old startup casts itself as an AI-safety research lab. But rather than operating as a non-profit like OpenAI, the company operates as a business and has been increasingly raising funds to build its own AI models. 

Given the reasons that have been circulating about the departure of Altman from OpenAI, one of them being AI safety policies, Anthropic’s approach might actually align with what the old OpenAI board wanted, including Ilya Sutskever. Now, the deal is no more.