Tipping the Scales of Reality: How AI Research Exposes the Industry’s Billion-Dollar Illusion

In a recent report that has sent ripples through Silicon Valley’s meticulously cultivated reality distortion field, AI researchers have delivered what can only be described as an inconvenient truth: throwing obscene amounts of money at a problem doesn’t necessarily solve it. Who knew?

The Association for the Advancement of Artificial Intelligence’s survey of 475 AI researchers revealed that a staggering 76% believe “scaling up” current AI approaches is “unlikely” or “very unlikely” to achieve artificial general intelligence (AGI). This collective academic eye-roll comes at a particularly awkward moment for tech giants currently building nuclear-powered monuments to their AGI ambitions.

Technology companies have spent years convincing investors, the public, and perhaps themselves that AGI – the holy grail of human-level artificial intelligence – is just a matter of more: more data centers, more processing power, more electricity, more billions. Microsoft alone has committed to spending $80 billion on AI infrastructure in 2025, an amount that could solve numerous pressing global problems but will instead be sacrificed at the altar of corporate techno-optimism.

The scaling dogma has always had a beautiful simplicity to it: if my model is smarter with 100 billion parameters, imagine how brilliant it will be with 500 billion! This logic, reminiscent of a teenager convinced their basketball skills would improve exponentially if only they had more expensive shoes, has driven investment strategies across the industry.

As Stuart Russell, computer scientist at UC Berkeley, eloquently puts it: “The vast investments in scaling, unaccompanied by any comparable efforts to understand what was going on, always seemed to me to be misplaced.” Translation: perhaps we should have spent some time understanding the technology before building data centers the size of small countries to power it.

Plateau of Diminishing Returns

Signs that the emperor might be underdressed have been emerging for some time. When OpenAI researchers discovered their next GPT iteration showed little improvement over its predecessor, one might have expected a moment of reflection. Instead, the industry doubled down, with Google CEO Sundar Pichai asserting there was no reason they “couldn’t just keep scaling up,” a statement with the same energy as a gambler convinced the next hand will definitely recover all previous losses.

Meanwhile, Chinese startup DeepSeek managed to create an AI model comparable to Western flagships at a fraction of the cost, using a “mixture of experts” approach that leverages multiple specialized neural networks rather than a single massive one. This is the algorithmic equivalent of realizing that assembling a team of specialists might be more effective than training one person to be mediocre at everything – a concept apparently revolutionary in AI development.

The Nuclear-Powered Elephant in the Room

Perhaps the most darkly humorous aspect of this scaling obsession is the energy consumption. Tech giants are literally signing deals to redirect entire nuclear power plants to fuel their data centers. One can’t help but picture future historians documenting this era: “And then, faced with climate crisis and energy insecurity, they decided to use nuclear energy to power machines that could write slightly better marketing copy and generate images of cats wearing hats.”

The irony thickens when considering that 80% of survey respondents believe current perceptions of AI capabilities don’t match reality. As Thomas Dietterich of Oregon State University notes, systems proclaimed to match human performance “still make bone-headed mistakes.” Yet somehow, these obvious limitations haven’t deterred companies from investing sums that would make the GDP of small nations blush.

Defining Success Through Moving Goalposts

Even the definition of AGI remains conveniently fluid. Google DeepMind describes it as a system outperforming humans on cognitive tests. Huawei suggests it requires a physical body. Perhaps most tellingly, Microsoft and OpenAI reportedly consider AGI achieved only when OpenAI develops a model generating $100 billion in profit – a definition that subtly shifts the goal from “human-level intelligence” to “unprecedented corporate profit machine.”

This definitional flexibility ensures the AGI carrot can always remain tantalizingly out of reach, justifying ever-increasing investments while providing a ready-made excuse for why true AGI hasn’t been achieved yet. “We just need more scale” becomes the technological equivalent of “the check is in the mail”.

The Scaling Bubble

As tech titans continue pouring billions into data centers that may well be elaborate monuments to a fundamentally flawed approach, smaller companies explore alternatives that do more with less. The situation bears an uncanny resemblance to other historical bubbles, where massive investment flowed into ventures based more on hype than substance.

Perhaps the most biting irony is that while AI models struggle to achieve genuine understanding, they have perfectly replicated one quintessentially human trait: the stubborn refusal to admit when a chosen path might be fundamentally wrong.

In this light, the tech industry’s continued commitment to scaling might represent the most expensive example of sunk cost fallacy in human history – a cognitive bias that even their own AI systems could probably identify, if only they were asked the right question.

Robert Nogacki

Robert Nogacki is a Polish attorney at law (radca prawny), the founder and managing partner of Kancelaria Prawna Skarbiec (Skarbiec Law Firm), which has operated continuously since 2006.

The law is equal for everyone, but the parties rarely are: on one side stands an organization with time, money, and lawyers, on the other a person with one business, one nest egg, and one life.

Clients rarely come to him with a legal problem. They come with a problem that also has a legal side: an audit that began with a single invoice, money entrusted to someone who has disappeared, a company that has to be passed on before it is too late. Most such matters are decided long before the first letter is written, in decisions made without asking and in deadlines nobody remembered. So he begins by asking how the client got here, not what the client should have done.

He advises entrepreneurs and families from more than a dozen countries, including those whose accounts the tax office has just seized and who do not know what to do tomorrow morning. He defends them in tax audits, customs and fiscal inspections, disputes with the tax authorities, and criminal tax proceedings. He represents victims of investment fraud and Ponzi schemes. He helps families set up family foundations and plan succession, so that a life’s work outlasts a single generation.

Not every case can be won. Every case can be run so that the client knows where they stand. Since 2006 he has represented the victims in the WGI case (Warszawska Grupa Inwestycyjna, the Warsaw Investment Group), one of the longest criminal cases in the history of the Polish financial market, because some things must not be left half finished, even when they take two decades. In the case of the collapsed cryptocurrency exchange Zonda (Zondacrypto, operated by BB Trade Estonia OÜ), he represents several hundred victims in the criminal investigation conducted by Poland’s National Prosecutor’s Office and in the Estonian bankruptcy proceedings.

Kancelaria Prawna Skarbiec is listed in the rankings of Poland’s largest tax advisory firms published by Dziennik Gazeta Prawna and Rzeczpospolita, and it is a four-time recipient (2015 to 2018) of the European Medal awarded by the Business Centre Club and the European Economic and Social Committee. Robert Nogacki publishes regularly, in the press and on the firm’s website, for people who have a problem rather than a law degree, because a legal opinion the client cannot understand protects only the lawyer.

He believes that the best legal advice is the kind that means the client never has to appear in court.