Parmy Olson (Macmillan Business (UK); St Martin’s Press (US))
For most people, ChatGPT appeared to materalise out of thin air. Within weeks of OpenAI’s quiet launch of the AI chatbot, it had become the fastest-growing app of all time and, almost two years later, it is nearly as well known as Google or Facebook. In the meantime, companies worldwide have gone gaga for the technology, with little time to pause to consider the wider societal consequences. So how did we get here and who was responsible?
Much of the praise (or blame, depending on how you feel about AI) lies with two men, argues journalist Parmy Olson in her book Supremacy: AI, ChatGPT and the race that will change the world. Sam Altman and Demis Hassabis, respectively the CEOs of OpenAI and London-based DeepMind, became involved with AI in very different ways, but their stories quickly became intertwined.
Advertisement
Altman, a Silicon Valley boy wonder, rose through the ranks of California’s start-up scene before setting his sights on what he viewed as the world-changing potential of AI. Hassabis, a child chess prodigy who flitted between the British video game industry and academia, was drawn to AI for different reasons, seeing it as a way to uncover the secrets of the universe.
Despite these differences, they found themselves in similar positions, steering AI research labs towards the ill-defined goal of “artificial general intelligence” (AGI). Both men started out with lofty ethical principles, promising to shepherd their AI systems for the public good and keep them out of the hands of profit-driven big tech companies. Both ended up financially entangled with those same companies, with Microsoft’s backing of OpenAI and Google’s purchase of Deepmind.
It’s a riveting tale and Olson is a compelling storyteller, drawing on the high-level access she has to both men, along with other figures in the story, thanks to her job as a tech columnist at Bloomberg Opinion. Neither Altman nor Hassabis comes off too badly in Olson’s telling, and she clearly has some sympathy for the complex web of philosophical and political movements they find themselves in as part of the AI field, like effective altruism, the “earn to give” ideology that disgraced crypto mogul Sam Bankman-Fried was famously a fan of.
One of the perils of writing about AI is that the tech progresses so fast that you risk becoming outdated
But this isn’t a hagiography either, and Olson is clear about the dangers that OpenAI and Deepmind have, in her view, conveniently ignored in an effort to push their technology out to the world so quickly and be first to claim the glory. Many of these will sound familiar – “hallucinating” made-up information, propagating the bias of the data they were trained on, generating offensive content – but Olson gives prominent voice to the researchers, nearly all women, who first raised many of these issues, and illustrates well how they were nearly universally ignored by their bosses, who were mostly men.
One of the perils of writing about AI is that the technology progresses so fast that anything you write is at risk of becoming instantly outdated. Olson largely avoids this by focusing on the story before ChatGPT, but there are times she must make an attempt to be as up-to-date as possible. Towards the end of the book, she covers more recent developments, such as the failed ousting of Altman by the OpenAI board last November, which still feels exciting, and the launch of tools like image generators and coding assistants, which feels less interesting and tangential to the earlier narrative of the book.
The story that has developed since Olson penned her last words, in March this year, has also been one of growing scepticism around whether AI models may have begun to approach a wall. OpenAI’s long-anticipated GPT-5 model is seen as a make-or-break moment for the industry, and whether it significantly moves the needle towards AGI, as either OpenAI or Deepmind have defined it, will dictate to a large extent whether the AI bubble will keep growing.
The subtitle of Olson’s book, “the race that will change the world”, makes that scenario sound like a certainty. But there is still a chance that our AI models might not get much more advanced than they are today, and these past two years will seem like an anomaly, rather than the norm.
The art and science of writing science fiction
Take your science fiction writing into a new dimension during this weekend devoted to building new worlds and new works of art
