The large artificial intelligence manufacturers wash away the negative image accumulated by the escapades of their models in recent months with an unusual wave of generosity. They offer products from the previously reviled open source and announce large public donations.
Billionaire Elon Musk and Dario Amodei, CEO of Anthropic, promise to give shares of their companies to children or charities and Sam Altman, of OpenAI, has offered to give a package of shares to the Government.
Additionally, Altman and Mark Zuckerberg, owner of Meta, embrace open source for some of their AI. Until now the main barons of artificial intelligence considered it a risk for their business. In summer they launched a crusade precisely against that distribution model. But they have not convinced the market and, on the contrary, “open source” AI made in the USA is increasing.
Meta, which launched two premium agents this month to compete with OpenAI and Anthropic, has teamed up with Apple so that an open version can be downloaded on that firm’s computers. It can thus be used autonomously and with less dependence on data centers.
The open code system or “weights” has been China’s battering ram to become the second power in artificial intelligence. Instead of residing on closed servers, and paying for access, users can save the parameters that feed the AI on their computers.
The open source of artificial intelligence has been reinforced with Reflection AI, a company founded two years ago in the US and already valued at 25 billion dollars. This month it launched Beam, the first cutting-edge American AI to adopt open weights as its flag.