A real countdown, which brought a package directly from Sam Altman under the tree. From December 4th to Christmas Eve, OpenAI accompanied users with an uninterrupted barrage of announcements and technological news. And the grand finale was the launch of o3 and o3-mini, the most powerful artificial intelligence models ever released by the San Francisco company. These models stand out for their abilities in solving programming and mathematical problems, achieving record scores in the ARC-AGI benchmark, a test designed to measure an AI's ability to autonomously acquire new skills.
With the release of o3 and o3-mini, OpenAI takes a significant step towards artificial general intelligence, a strategic goal pursued by Altman. AGI is conceived as a system capable of performing non-specifically programmed tasks, reasoning similarly to a human being. Initial access to these new models will be restricted to a select group of researchers, but o3-mini is expected to be available to the public as early as the end of January.
The most commercial change in strategy
The announcement marks a further evolution towards a more commercial and marketing-oriented strategy for OpenAI. Last week, the company gave 2.7 billion WhatsApp users free access to ChatGPT, opening up new usage opportunities for millions of people. Again with a view to broadening its audience, it also launched the SearchGPT search engine, available for free in Italy, with the aim of challenging Google and bringing AI to the center of online searches.
On December 10, OpenAI also presented Sora, a platform that allows users to generate videos that are virtually indistinguishable from reality. This development represents a further example of the power and versatility of artificial intelligence technologies, but also raises concerns about the authenticity of digital content. So, to counter these risks, Meta recently launched Video Seal, a technology that adds a watermark to AI-generated videos to verify their authenticity and stem the spread of deepfakes. And above all it unleashed a counteroffensive against its great rival.
Legal disputes
In fact, the more OpenAI broadens its horizons, the more the competition tries to block its developments. Elon Musk (who was a co-founder of Altman's company before leaving in disagreement with AI developments) and Mark Zuckerberg, CEO of Meta, have asked the California attorney general to block the transformation of OpenAI into a for-profit company. According to Zuckerberg, this change could set a dangerous precedent for the entire Silicon Valley, allowing anyone to found a non-profit, enjoy tax advantages and then transform into a commercial company without losing initial investors: if OpenAI's transition were to be completed , the company would save billions of dollars, strengthening its dominant position in the artificial intelligence sector, already valued at $157 billion.
The prospect of an OpenAI monopoly, supported by the latest record financing of 6.6 billion dollars, therefore worries competitors, also because nVidia and Microsoft stand out among the main investors, with the latter having already invested around 14 billion dollars in Altman's company.
This shows that the race towards general artificial intelligence is in practice more crowded than ever, and even if OpenAI remains firmly in the lead, the question is whether it will manage to maintain the balance between technological progress, economic sustainability and ethical responsibility. The answer, for some, may not be comforting.
