I'll be honest: Amazon's announcement of its investment in Anthropic, another pioneer of language models, marks a new chapter in the AI industry but it doesn't surprise me. In my mouth everything tastes like inevitability.
In the beginning it was Google, which moved well ahead of the others (the purchase of Deepmind it even dates back to 2014). Then the “sudden” advent of OpenAI since 2019, which has accelerated the commercial use of these systems with the now well-known ChatGPT. Next, the Microsoft's first move which focused on OpenAI and Google's confused reaction (before Bard, then Gemini & co.), a giant taken by surprise. Elon Musk is not missing from the roll call, working somewhat quietly on X.ai and Dojo supercomputer. Could Jeff Bezos have been missing?
Computing power
Amazon is the company that transformed the way we shop online. It established an almost monopolistic position, extending its activities into the field as well of shipments, consumer electronics, clothing.
In terms of AI, to be honest, it didn't particularly shine (especially in the first versions of its voice assistants). Yesterday it announced an investment of up to $4 billion in anthropic, an emerging company in the Artificial Intelligence arena. But who is Anthropic? They are known for developing language models such as Claudia 2, which stands as a direct competitor to OpenAI's ChatGPT.
It's not just about putting money into it. With this partnership, Amazon it also becomes the primary supplier of computing power, or “compute,” for Anthropic. If you're wondering why this matters, think about an investor who buys a small airline and provides it with infinite money, infrastructure, and even fuel. AI models require enormous amounts of computing power to train and run. And this need will grow exponentially.
This partnership not only strengthens Amazon's position in the industry, but could also redefine the future of AI. What are the implications of this collaboration for the technology ecosystem and consumers? I try to understand.
Anthropic and Amazon, win-win
Clearly Amazon is aiming for a large profit. In exchange for its substantial investment, it will acquire a position (albeit a minority one) in Anthropic. This means Amazon engineers will be able to integrate Anthropic's AI models into their products and services. Alexa, starting today you will no longer be a simple pasta timer.
Founded in 2021 by a group of former OpenAI employees, Anthropic has quickly gained traction as a leading foundation model developer. But what drove these employees to leave OpenAI? Differences in their approach to AI safety. And now, with Amazon's backing, they're ready to take their vision to new heights.
With a few more compromises: the increase in investments by cloud computing providers will allow Big Tech companies to exercise more and more control over AI developers. Your ethics are fine, Anthropic, your talent is fine, but when your partner puts the "fuel" to make you move (in this case the computing power) you could put aside a lot of idealism.
Competition concerns
Not everyone sees these partnerships in a positive light. Both the UK's antitrust regulator and the US Federal Trade Commission have expressed concerns. These partnerships could allow large companies to extend their monopoly power, influencing the entire AI ecosystem (and a possible piece of the future, as we have seen recently).
Certainly, with giants like Amazon entering the game, the AI landscape is destined to change, and protection bodies would do well to hurry: the "war" has just begun.
