Sam Altman, CEO of OpenAI, is adopting the “move fast and break things” strategy popularized by Mark Zuckerberg in the early days of Facebook. In fact, he has been doing it for years and it is even confirmed by his own tweets from the past, but in recent weeks it has become especially evident.
What happen. On the one hand, he has apologized for several decisions in recent months, such as the unauthorized use of a voice intentionally similar to that of Scarlett Johansson or the abusive clauses in his employees’ contracts that he said he was unaware of (he signed them). On the other hand, he continues to accelerate with the development of GPT-5, his next LLM.
It’s an approach that rewards speed over caution, but with AI comes with potential risks.
Because it is important. AI has transformative, but also destructive, potential. It is not necessary to be catastrophic, but it is not healthy to ignore risks either. As happened with social networks fifteen years ago, but at a much more dangerous level in the long term.
Repeating the strategy that led Facebook to connect the world at the cost of polarizing and misinforming it can have dire consequences if applied to the development of AI. What did you say Axios: The industry continues to reward those who ask for forgiveness instead of permission.
Altman’s philosophy. In 2021 she tweeted “Move faster. Slowness anywhere justifies slowness everywhere.” Another aphorism signed by him was “Momentum is everything in a startup. Those who win keep winning, those who lose keep losing.”
- Being in charge of the OpenAI security committee is the AI equivalent of putting the fox in charge of the henhouse, said with love.
- Both CEOs share both optimism about the positive potential of their companies and deliberate naivety about their externalities.
Maneuvering Zuckerberg style. Mark has been apologizing for Facebook/Meta’s errors and inaccuracies since its inception. but he has never stopped his plans. He did it with him feed news feeds, privacy settings, fake newsthe Cambridge Analytica scandal, child abuse…
Facebook and Instagram always continued to grow in users, engagement and benefits. Altman is repeating the pattern: he apologizes for the controversies while presenting new demos, announcing the development of GPT-5 and liquidating the team that controls the security of that development.
The risks of haste. From the hallucinations that we have already become accustomed to to some grotesque first versions. The start of Bing’s GPT-based AI was also a farce.
The generative AI race, especially between Microsoft, OpenAI and Google, is leaving us with products that seek to impact and whose risks and inaccuracies we will discover as we go. All for the sake of getting ahead.
Lessons not learned. Zuckerberg left his old mantra for dead after the Cambridge Analytica scandal, but that mantra, although now technology companies are so big that they have become mammoth and more cautious, is still alive in the industry.
The big problem: the disruption of AI can have a much greater and more damaging scope than what has been seen with social networks in the last three decades. This approach may give OpenAI a competitive advantage, but it runs the risk of repeating or compounding the big mistakes of the golden age of social media.
In Xataka | OpenAI is clear about how to make ChatGPT Plus succeed: using ChatGPT as bait
Featured image | Xataka with Midjourney
