Ed Zitron Discusses Big Tech, Backlash, Boom and Bust: ‘AI Shows People Are Eager to Replace Humans’

If a film is made in the future about the rise and fall of the AI bubble, Ed Zitron will undoubtedly emerge as a key figure. He epitomizes the outsider perspective: the quirky, solitary individual who recognized the impending chaos and loudly proclaimed the crisis, only to be ignored. Much like Christian Bale’s performance as Michael Burry, the shrewd investor who foresaw the 2008 financial collapse in The Big Short, one can easily envision Robert Pattinson or Paul Mescal representing Zitron, the fervent, vividly expressive, but meticulously detail-oriented critic of big tech.

While it’s uncertain if the AI bubble will indeed burst, Zitron’s unapologetic skepticism amid a sea of AI enthusiasts has made him somewhat of a cult figure. His tech newsletter, Where’s Your Ed At, boasts over 80,000 subscribers, while his weekly podcast, Better Offline, consistently ranks among the Top 20 in tech charts. A frequent contrarian voice in media, his subreddit has evolved into a refuge for AI skeptics, even attracting insights from those within the tech industry. One poster famously referred to him as “a lighthouse in a storm of insane hypercapitalist bullshit.”

Zitron’s exploration of generative AI began in 2023, following the groundbreaking launch of OpenAI’s ChatGPT. “The more I delved, the more perplexed I became because, in addition to the fact that large language models (LLMs) clearly didn’t achieve the anticipated feats, they also lacked any plausible route to doing so,” he explained. “Nothing I discovered indicated this was even a viable business, much less a world-altering innovation.”

Speaking via video call from his Las Vegas office, Zitron, clad in a red hoodie and surrounded by framed pop culture artwork, exhibits an impressive ability to articulate his views. Those familiar with Better Offline know he is a master of monologues, delivering his perspective in straightforward, humor-laced language filled with data, anecdotes, and occasional expletives, all delivered in a distinctive London accent that emphasizes his role as a Silicon Valley contrarian who playfully drops his Ts when he mentions “datacentres.”

Distilling Zitron’s critique of generative AI is no simple task: last year, he penned a thorough 19,000-word essay presenting his case. However, it can be summarized into two interconnected aspects: the real-world effectiveness of the technology and the underlying financial framework supporting the AI boom. Zitron believes that both bases are fundamentally unstable.

The first concern centers on whether generative AI can deliver on its lofty promises. The narrative around the technology has been rife with predictions of its ability to obliterate conventional employment structures. Dario Amodei, CEO of Anthropic—the main competitor of OpenAI—cautioned last May that AI might eliminate half of all entry-level white-collar jobs within five years. Zitron, however, refutes this with conviction. “The current generation of LLMs will not achieve that. My proof is that their capabilities have remained consistent over the last year. They’ve not found a way to perform tasks autonomously, and every test to enhance their functionality seems to have failed.” He highlights that LLMs produce erroneous content, offer inconsistent responses, and struggle with complex tasks, raising doubts about even labeling this technology as “intelligent.”

“It’s as intelligent as a pair of dice,” he quips. “LLMs operate through transformer-based architecture utilizing vast probabilities to generate outputs. They do this on a grand scale, leading one to think they are creating content. However, they rely on a massive database and numerous parameters to produce any result. That’s all there is to it. We wouldn’t call an Excel formula intelligent, so we shouldn’t attribute that status to generative AI.”

Many individuals contest Zitron’s views, especially regarding AI’s potential to take jobs. Across various sectors—film, customer service, government, and tech—insiders argue that AI tools empower them to accomplish tasks with fewer resources. Even if AI doesn’t eradicate 50% of jobs, its transformative impact on the workforce is probable. A survey conducted last June revealed that entry-level roles had declined by nearly a third in the UK since ChatGPT’s launch.

Zitron counters that “correlation does not imply causation,” referencing findings that substantiate the assertion that machine learning’s role in job reductions may be exaggerated. A recent MIT report on the “state of AI in business in 2025” indicated that 95% of firms attempting AI integration were getting “zero returns.” “Most GenAI frameworks do not retain feedback, adjust to users’ contexts, or evolve over time,” the report noted.

Moving on to the second component of Zitron’s argument: the economics behind the AI explosion are alarming. The sheer volume of investments flooding into AI is unprecedented. The “magnificent seven”—Alphabet (Google’s parent), Amazon, Apple, Meta, Microsoft (which owns 27% of OpenAI), Nvidia, and Tesla—currently account for 34% of the S&P 500, the US stock market index representing about half of the global market. Nvidia, the leading GPU manufacturer that powers AI operations, is “basically printing money,” according to Zitron. Yet, many other firms are incurring debts and spending billions without the prospect of recovery.

Historically, Silicon Valley startups have operated at initial losses to gain market share in hopes of future profits. However, the current imbalance between supply and demand is dramatically concerning. Building a robust AI platform necessitates significant financial investments. A typical datacentre comprises tens of thousands of GPUs, each priced at over $50,000 (£37,000). Additional costs include specialized software, networking solutions, massive facilities, and substantial energy usage. The cost for constructing 1GW of AI datacentre capacity is estimated at $35 billion (£26 billion). Consequently, only financially robust companies—referred to as “hyperscalers,” such as Google, Meta, Amazon, Microsoft, and Oracle—are equipped to engage in this market.

Yet, from the demand perspective, the outlook appears grim. OpenAI has pledged to invest $1.4 trillion (£1 trillion) in AI infrastructure over the next five years, with projected revenues of merely around $20 billion (£15.8 billion) for 2025. There seems to be a constant flurry of deals between AI entities, but upon scrutiny, many transactions essentially involve companies transacting with each other. For example, Nvidia recently announced a $100 billion investment in OpenAI, while OpenAI will use those funds to purchase Nvidia chips. Similar contingent arrangements are prevalent within the industry, as Zitron meticulously outlines. Even lesser-known “neocloud” providers, such as CoreWeave, Lambda, and Nebius, primarily derive their business from the larger players like Google and Microsoft. Zitron argues, “Subtract the hyperscalers, and the entire AI compute sector is projected to generate less than a billion dollars in 2025.”

Regarding profitability, while ChatGPT reportedly has around 800 million users, most have not paid for the service. Even those who do pay face variability in costs associated with interacting with the AI model. “Each query’s cost fluctuates significantly based on complexity,” Zitron notes. This lack of economies of scale means that each interaction incurs substantial expenses for providers. “The more vigorously people use these platforms, the higher costs will be for them,” he warns. If users request adjustments, which is often necessary due to model inaccuracies, this adds more computational load without generating additional revenue. Despite claims of reducing operational costs and refining AI systems, this usually entails greater consumption of computing resources. “It’s akin to decreasing petrol prices but requiring an additional 250 miles of travel to reach a destination. Hence, it’s a significant concern, suggesting that profitability may be elusive.”

Although none of this guarantees an imminent crash of the AI sector, Zitron expresses uncertainty: “If I am mistaken, I can’t fathom how or why.” He mentions that the counterarguments he’s encountered largely seem to involve mere optimism that “AI will improve over time.”

Critics often accuse Zitron of harboring an agenda against big tech, yet he challenges this notion: “My actual concern is with those who refuse to confront reality.” While he embraces scrutiny, he insists it isn’t the primary motivation behind his careers, stating: “I enjoy writing. I relish dissecting issues. I find joy in solving puzzles. A considerable portion of my work is driven by a desire to comprehend things myself rather than merely informing an audience.” Zitron doesn’t possess formal training in economics or computer science and has never held a tech job. “I’ve taught myself through experience.”

Nonetheless, Zitron has always been inclined toward technology. He recalls building ten personal computers throughout his life, sparked by his father gifting him a dial-up connection card at age ten. “From that moment, I was online, and I thought, ‘This is the future. I love this ability to connect and play games.’ I was quite a lonely child, but I managed to forge many friendships online.”

Growing up in Hammersmith, London, Zitron describes his parents as supportive. His father worked as a management consultant, while his mother nurtured him and his three older siblings. However, he struggles with his experiences in secondary school, vaguely mentioning past difficulties. Diagnosed with dyspraxia and ADHD, he candidly admits to struggling in the traditional education system. “I believe I failed most languages and science courses, and my math wasn’t stellar either,” he reflects. “However, my focus has always been meticulous.”

After earning a degree in media and communications from Aberystwyth University, Zitron began writing for gaming magazines. Eventually, he reached a point of discontent in London and decided to move to New York in 2008, where he ventured into tech PR. He has no intention of returning to the UK, stating, “I don’t speak about my personal life much but I do have a son, which is why I relocated to Las Vegas. I enjoy the oddity of the place: ‘Everyone’s strange, so no one’s strange.’ It seems he has been previously married and divorced twice.

Zitron remains entrenched in tech PR, a role that seems contradictory to his position as a tech dissenter—akin to biting the hand that feeds him or raising a conflict of interest. He doesn’t see it that way, insisting he has no AI clients or engagement with major tech firms, stating he currently has only a few clients. This involvement has equipped him with a network of industry contacts, potentially boosting his self-promotion (he published a book in 2013 titled This Is How You Pitch: How To Kick Ass in Your First Years of PR). However, as his media presence increases, it may eventually lead him to pivot away from PR. He’s working on a new book scheduled for release next year titled Why Everything Stopped Working. “It’s a deep dive into how society has evolved and why technology is now indispensable,” he adds, indicating that a chapter will specifically address AI.

If Zitron has other axes to grind, they align with broader critiques of neoliberal capitalism. “I don’t believe people grasp the detrimental consequences of deregulating financial markets as promoted by Thatcher and Reagan. The lack of accountability after the 2008 financial crisis is often overlooked, and the risks associated with growth-focused capitalism are severely underestimated.”

Zitron suggests that rather than ushering in an era of prosperity, AI personifies the culmination of neoliberal ideals. “The most telling lesson from the latest generation of language models is how eager people are to replace human workers, often without understanding the value of human labor,” he observes.

Zitron isn’t alone in his assessments any longer. His views resonate with contemporaries such as Cory Doctorow, who has appeared on his podcast and posits the “enshittification” paradigm that asserts tech firms now prioritize profit over useful innovations. Various other AI skeptics, like cognitive scientist Gary Marcus, remark they have championed similar positions as Zitron’s but feel overlooked within his narrative. Concurrently, the backlash against AI is intensifying: local organizations are opposing the unveiling of environmentally detrimental datacentres; consumers are pushing back against the algorithmic invasion of products; and creators are pursuing legal recourse against industries profiting from their work. Public outrage continues to mount over social media-related harms, epitomized by Elon Musk’s Grok producing nonconsensual borderline-deepfake content.

Speculation surrounding the AI bubble’s potential collapse continues to gain traction. Figures from the Bank of England to Microsoft’s CEO Satya Nadella are sounding alarms. Michael Burry, known for his predictions in the financial crisis, has expressed he is shorting Nvidia. Recently, the New York Times published an opinion piece suggesting that OpenAI is at risk of financial insolvency within the next 18 months. Zitron posits it could happen sooner, particularly with major tech companies ready to announce their annual earnings for 2025. They’ve been reticent about their AI-specific revenue, leading Zitron to question, “Why the secrecy? It indicates those figures must be quite modest. If something significant transpires, like Nvidia falling short of its projections, it might lead to a widespread reassessment of the entire sector, risking a new global economic crisis. Those datacentres might become vacant shells. We might be witnessing what he humorously labels, “the largest laser-tag arena construction project in history.”

Zitron doesn’t derive any pleasure from being a contrarian. “It’s not enjoyable being on the periphery of an idea. Many individuals favor the supportive narrative surrounding AI as it’s undeniably easier.”

He doesn’t harbor animosity towards tech, or AI for that matter. “I have a passion for technology, but I disapprove of how the tech industry operates… If questioning these developments is branded as anti-innovation, it highlights the unequal power dynamics we currently face. It creates a scenario in which even affluent individuals feel compelled to acquiesce to these corporations. Yet, these companies have contributed minimally to enhancing our daily lives while amassing vast wealth that surpasses our own.”

He simply desires to spotlight the underlying truth. “Sure, it would be simpler to craft mythologies and fan fantasies about the potential of AI. My goal is to decipher the reality of the situation.”