On AI Hype, Tech Trends, and the Tyranny of Mainstream Bias

By Dr. Hung-Chih Lu

The past decade-plus was the golden age of software. Software innovation led the charge in technological development. But since AI emerged in 2022, we have entered a spiral of displacement in less than four years.

The 10-nanometer threshold divides advanced from mature semiconductor processes. AI is the equivalent divide in software: anything achievable through Vibe Coding belongs to the mature category. Just as mature-node chips have become commodity goods, so too have mature-category software products. Advanced processes and advanced software remain premium. The difference is that hardware, which once played second fiddle to software, has now reversed that relationship.

On top of this, the enormous demand for compute has kept advanced hardware firmly in the category of hard currency, while advanced software companies are hemorrhaging money. Apple chose not to pour capital into building its own AI; instead it brought in Gemini. The result: Apple's financials have been remarkably strong, putting pure-play AI software companies to shame by comparison. Google, meanwhile, uses Gemini to extend its reach into the iOS ecosystem while leveraging its own Android platform and cloud services in a combined hardware-software strategy. Its decision to open-source Gemma 4 for free has placed enormous pressure on both closed-source and budget open-source AI model companies.

Large AI model companies, operating in a state of deeply imbalanced investment-to-return ratios, are locked in a game of chicken. The capital they must deploy keeps growing, yet they face stubborn problems: high AI hallucination rates, staggering compute costs, and revenue that cannot translate into real profit. They need massive compute infrastructure to survive, but they cannot afford to sustain it. It is a true dilemma.

The current AI supply chain works like this: a farmer grows vegetables (AI hardware) and sells them for 10 units each to a market stall (AI software). The stall sells them to consumers for 5 units, losing 5 units on every sale. Because consumer demand is strong, the stall commissions the farmer to grow more. The more the stall sells, the more it loses. The more it loses, the more it sells. It cycles deeper into negative margins, running through capital and then raising more rounds, all while the company's valuation climbs higher. How long can this possibly continue? When the stall can no longer find fresh capital, it collapses. The farmer is left with mountains of unsellable produce, even after cutting prices. That is what an AI bubble bursting looks like.

The winners in this landscape will be those who grow and sell their own produce and can regulate supply (the integrated hardware-software players, Google being the prime example), or those who operate as platform landlords collecting listing fees and rent (Apple being the prime example). Everyone else will be affected to varying degrees. Advanced hardware companies will see profits sharply corrected, but not collapse. Advanced software companies will undergo brutal consolidation. The integrated advanced hardware-software companies will stand on the most stable ground. The rest will become targets for investment groups to pick over at their leisure. Soft assets may lose their value, but hard assets remain highly attractive.

AI's impact is deep and wide. Where a task once required one hundred people, AI may reduce that to ten. Jensen Huang says that AI will transform the way people work rather than eliminate jobs. He is talking about those ten people who remain. He never mentions the ninety who disappear, in the same way a self-styled prophet proclaims the one prediction that came true and quietly forgets the nine that did not.

Consider a concrete example: a real estate tycoon who wanted the prestige of being called a tech mogul set up a technology company and hired hundreds of employees, content to break even. In the AI era, he would not need hundreds of employees. He would not need to learn AI himself. He would simply spend money to hire the best AI talent, and the total salary bill would still be lower. In the AI era, the winners are the wealthy, not those who are clever with AI tools. The rich find it easier than ever to make money from money.

Some economists invoke the spillover effect: if AI makes a small number of people extremely wealthy and they spend more, it might indirectly create some low-skill or service-sector jobs. But this is a reasoning error rooted in a minority mindset. The wealthy are, by definition, a small fraction of the population. No matter how much they eat and spend, a tiny number of people cannot substitute for the economic activity of the many. Not a single national economy in the world is sustained purely by the consumption of the rich.

No matter how those with vested interests try to shape mainstream narratives through every channel available to them, truth cannot be buried. Under capitalism, the greatest beneficiaries of the AI era are, and will remain, the small minority who were already wealthy.

In the name of truth and public good, I have pursued an entirely self-funded creative path spanning science, art, and literature. From my earliest work in 1999 on a Multi-Chip Integrated Silicon Photonics Architecture to my most recent work in 2024 on Distributed AI Robots and a Pixelated Modular Architectural System, I have produced over a hundred inventions and dozens of patents. Not a single one has been used for commercial gain. Before AI existed, my method was to file utility model patent applications, then let them lapse after registration so the inventions would enter the public domain as quickly as possible. Now that AI exists, simply making something public means any subsequent patent application can be challenged for lack of novelty. The work immediately becomes public property the moment it is disclosed, so I no longer need the old method.

Funding this research independently, so as not to be steered by investors, has put me through no small degree of hardship. To keep costs down, I have handled everything myself, large and small. My income has come mainly from technical consulting, web design, graphic design, and custom poetry writing, each corresponding to the domains of technology, art, and literature. These services aligned naturally with my own values of public benefit. Since AI arrived, those commissions have dried up entirely. Funding is now scarce, but my work has not stopped.

Because I spend my own money, I answer to no mainstream value system and no market economy. The cost of that freedom is a fate not unlike that of Nicolaus Copernicus. When he proposed the heliocentric model, and Galileo Galilei later corroborated it, they were challenging the most powerful authority of their time, the Church, as well as the evidence of people's own senses. When Nikola Tesla championed alternating current, he faced the full-force suppression of Thomas Edison, who held the monopoly on direct current. What these pioneers shared was the courage to stand against mainstream consensus, indifferent to personal glory, survival, or loss, committed only to truth and the public good. The mainstream values of any given era are almost always built on the inertia of past success. Truth is the only exit toward the future.

Whether it is Wikipedia or the Nobel Prize, recognition flows toward those who have earned mainstream academic approval. When I proposed my Multi-Chip Integrated Silicon Photonics Architecture, I was mocked as an amateur. There was never any possibility that the academic establishment would validate me. To dismiss me by pointing to the absence of mainstream recognition, rather than examining the actual evidence, is absurd. In the era of floppy disks and CDs, much of the documentation was lost to data corruption. Fortunately, the critical evidence remains intact. Without it, I would have had no way to vindicate myself.

In terms of conventional returns, whether reputation or profit, I have come out deeply in the red. The world looks up to the tech billionaires who harvest its wealth, not to people like me. But measured in terms of public benefit, the return on investment is infinite.

People have always questioned me: the world does not appreciate what you do, so why keep giving? My answer is that most people in this world are good. You cannot judge the whole by a small part. Those who do not see it are simply blinded by mainstream bias. What I do is follow the path of those who came before me, holding to the values of truth and public good. Even if I stand alone against ten thousand, I will walk forward. The credit need not be mine. The Nobel Prize will never come to me. But if the world is made better, I have no regrets.

postal code: 328015, Taoyuan city, Taiwan

Dr. Lu's Studio
Devoted to the public welfare creations of ecology, technology, and art.

This release was published on openPR.