The book opens by confronting the reader with a truth that now feels uncomfortably familiar: the line between human and machine-created content is disappearing fast. Parmy Olson chronicles how the release of ChatGPT in late 2022 turbocharged public fascination with AI, sparking awe, fear, investment booms, and cultural shock in equal measure.
Within just two years, AI went from a niche technology generating distorted dog pictures to tools capable of photorealistic images and eerily convincing conversations. For many, this shift felt like a technological ambush, upending everything from creative professions to perceptions of truth, agency, and even consciousness.
At the centre of this revolution are two men: Sam Altman, CEO of OpenAI, and Demis Hassabis, founder of DeepMind. Both began their journeys with idealistic missions to improve humanity—but the book deftly shows how capitalism, ego, and Big Tech incentives slowly warped those missions into something less utopian.
Sam Altman – The Silicon Valley Prodigy
Altman’s story is archetypal Silicon Valley. A Stanford dropout and founder of the social app Loopt, he proved himself a persuasive communicator and aggressive networker. His success with Y Combinator solidified his status as a startup kingmaker, but it was his growing obsession with Artificial General Intelligence (AGI)—AI that could outperform humans at virtually any task—that reshaped his destiny.
He envisioned AGI as a force that could end poverty and increase global prosperity, but he also sought control. Altman wasn’t just racing to build AGI—he was racing to define the narrative around it, to position OpenAI as a guardian of humanity.
Demis Hassabis – The Chess Genius Turned Neuroscientist
Hassabis, in contrast, was the brainy prodigy raised in North London, once the second-best chess player in the world under 14. After a failed stint in the gaming industry, he pivoted to neuroscience, obsessed with unlocking the mysteries of human cognition. This obsession led him to DeepMind, which would be acquired by Google and famously beat the world champion at Go.
While Altman believed AGI would create material wealth, Hassabis saw it as a path to scientific discovery—a tool to understand the universe, cure diseases, and perhaps even answer metaphysical questions about consciousness and the mind.
Their contrasting visions—abundance vs. enlightenment—would eventually converge into a high-stakes rivalry.
ChatGPT and the Beginning of the Race
The launch of ChatGPT was a turning point. It wasn’t just a product release; it was a cultural event. Suddenly, ordinary people were interacting with AI that felt clever, funny, even human. This mainstream moment created a scramble among Big Tech firms to launch competing products.
Google’s rushed release of Bard (based on LaMDA) and Microsoft’s integration of ChatGPT into Bing illustrated how existential the threat was perceived to be. Generative AI was now the new platform war—and both OpenAI and DeepMind were no longer independent labs; they were now deeply entangled with Microsoft and Google respectively.
The Faustian Bargain
Despite early claims of prioritising safety and transparency, both OpenAI and DeepMind were lured into classic Silicon Valley mission drift. Altman, once critical of profit-driven AI development, handed more control to Microsoft in exchange for billions. Hassabis, once focused on science, led a merger of DeepMind and Google Brain to stay competitive.
In both cases, access to data, computing power, and scale won out over independence and ethics. Transparency suffered, public scrutiny was evaded, and internal critics (like Timnit Gebru and Margaret Mitchell) were either silenced or marginalised.
Hype vs. Harm
As Altman toured Congress warning of the potential existential threat of AI, many critics saw a clever bait-and-switch. By focusing on speculative, distant dangers—Skynet scenarios and paperclip maximisers—OpenAI and its peers deflected attention from real, present harms: racial bias in image generation, misinformation, job displacement, and exploitative data labour.
AI ethics researchers—many women and people of colour—had long warned of these harms. But their budgets were minuscule compared to the billions pouring into “AI safety” groups fronted by elite men warning of distant doom.
Olson draws a powerful comparison to the plastic industry’s PR around recycling—a symbolic distraction from the unsustainable growth of plastic itself. In the same way, fear of killer robots allows AI giants to avoid regulating their current models.
Who Builds the Data? Who Gets the Power?
Behind every powerful AI model lies a hidden workforce—often low-paid data labourers in countries like the Philippines, Kenya, and Mexico. These workers label datasets, filter toxic content, and perform menial annotation tasks that form the backbone of tools like ChatGPT and Bard. Yet they remain invisible in AI marketing or discussions of innovation.
Meanwhile, regulators lag behind. A Stanford study cited in the book found “virtually no transparency” in how AI models are built, deployed, or monitored. Governments have no clear oversight, and the public is increasingly in the dark about how these tools shape our information and experiences.
The Altman-Hassabis Rivalry
Though both Altman and Hassabis began with philosophical motivations, their rivalry became tactical and corporate. Altman, ever the showman, leaned into doom-laced testimony and massive PR tours. Hassabis, more reclusive, was quietly trying to keep up, even as Google scrambled to maintain relevance.
But both were now essentially generals in a corporate proxy war—Altman for Microsoft, Hassabis for Google. As they competed to build smarter models and claim dominance, their visions became compromised by the very forces they once opposed.
Anthropic and the Splintering of the Field
Alongside OpenAI and DeepMind, a third player emerged: Anthropic, a company founded by ex-OpenAI researchers. Claiming a more safety-first approach, it too began accepting billions from Google and Amazon. The same safety rhetoric masked aggressive ambitions to dominate the LLM race and expand into dozens of industries.
Anthropic and others justified building powerful models by arguing that you can’t make them safe unless you build them first. Olson exposes this for what it is: a moral loophole used to justify exponential growth while downplaying risks.
From Surveillance to Cognitive Decline
The second half of the book delves into where this all leads. AI is being embedded in dating apps, entertainment, government welfare systems, and even policing. Generative AI can now produce personalised ads that say your name, or bots that flirt on your behalf.
Olson raises concerns about how this affects cognition and agency. Like the “Google Effect”—where constant internet use weakens memory—generative AI could degrade our problem-solving abilities. Coders already report struggling when GitHub Copilot goes offline.
Bias and Inequality on Steroids
Bias remains a massive concern. Ask some image generators for CEOs, and they give you white men. Ask for a criminal, and it’s often a Black man. These biases are no accident—they reflect the data used to train models. Worse still, AI-generated content is flooding the internet, making it harder to distinguish real from fake. Soon, 90% of all content online may be machine-generated.
Inequality could be supercharged. Wealthy users and firms will have the best AI tools; others may get flawed, surveilled versions. AGI, if it ever arrives, may not be a public good—it might be a paywalled service controlled by trillion-dollar firms.
The Illusion of Altruism
Olson closes with a sobering parallel: Altman and Hassabis’s journey is not unlike the story of Edison and Westinghouse, who fought to electrify the world. But it was General Electric—representing corporate consolidation—that ultimately won.
Both Altman and Hassabis, despite their idealism, ended up handing over the reins to Big Tech in exchange for resources, distribution, and prestige. The result is a world where AI is poised to reshape humanity, but the steering wheel is in the hands of a few powerful firms—not democratically accountable institutions.
Their intentions were noble. But their tools, and the forces around them, were stronger. And now the world is left to deal with the consequences of their race for AI supremacy.
Do You Want To Learn Scrum & Agile?
Learn Scrum & Agile at lost cost and online. Our 5-star rated Ultimate Scrum & Agile eLearning Courses can take you from beginner to advanced at your own pace.
Prepare and practice for the assessments from the major Scrum & Agile providers. Our 5-star rated Ultimate Scrum & Agile Practice Assessments will help you gain certification.
About TheScrumMaster.co.uk
