By Byron V. Acohido
Last Friday morning, April 11, I was making my way home from NTT Research’s Upgrade 2025 innovation conference in San Francisco, when it struck me that we’re at a watershed moment.
Related: How GenAI is disrupting the value of legal work
I was reflecting on NTT’s newly launched Physics of Artificial Intelligence Lab when a GeekWire article crossed my LinkedIn feed, touting a seemingly parallel initiative by Amazon. But while the surface resemblance is easy to draw, the underlying intent—and trajectory—sets the two efforts worlds apart.
Amazon CEO Andy Jassy had just publicly anointed Alexa+ as, in effect, the spiritual successor to Gutenberg’s printing press. In his annual letter to shareholders, Jassy presented Alexa+ as the first personal assistant that can truly act, declaring “Generative AI is going to reinvent virtually every customer experience we know.”
Contrast that with how NTT Research has set out to define the mathematical underpinnings of language-based AI through physics, neuroscience, and moral psychology. On one hand, NTT Research CEO Kazu Gomi drew this analogy: for millennia, humans watched apples fall without understanding why. It wasn’t until Newton came along that we could calculate the immutable forces—mass, distance, rotation—that govern motion.
Today, the output of generative AI feels just as forceful, yet far less understood. It shifts in response to invisible factors—training data, model weights, prompt context—but we don’t yet have the math to explain it. Gomi’s point: we’re living in a pre-Newtonian era of AI. And if we hope to guide it, we’ll need new laws—new physics—for cognition itself. Like a haiku, his message was spare but profound: clarity precedes control.
On the other hand, Jassy asserted Amazon’s positioning of its custom chips (Trainium2), foundation models (Nova), and hyper-integrated stack (Bedrock, SageMaker) not just as tools—but as rails upon which all future AI must run. It’s a powerful story, delivered with discipline—but we’ve seen this movie before.
What’s in the black box?
This isn’t just about product launches and shareholder letters. What I witnessed at NTT’s Upgrade 2025 was something altogether different: a quiet but profound attempt to chart a moral and scientific path through the black box of AI. While Amazon rushes to own the rails, NTT is asking: what is this machine we’ve created—and what kind of citizen might it become?
Take a moment to connect the dots. From Gutenberg’s press to the steam engine, to the rise of semiconductors—each transformative leap began as an open revolution and was soon constrained by consolidation. Bezos launching Amazon with a single book, and Google’s Brain Team engineering the transformer architecture that underpins today’s GenAI—these are milestones on the same arc. Now return to Gutenberg: the deeper lesson isn’t just invention, but how control followed innovation.
Co-opting innovation
Gutenberg’s movable type press, invented around 1440, is widely credited with kickstarting Europe’s cultural and intellectual awakening. But the full flowering of its revolutionary impact didn’t unfold until centuries later, when literacy became accessible to commoners, not just clergy and nobles.
What happened in the interim? The early printing industry was co-opted. Powerful institutions—the Catholic Church, monarchies, and merchant guilds—quickly moved to control access, censor dissidents, and monopolize distribution.
Fast-forward 560 years. Jeff Bezos sells his first book on a novelty called Amazon.com in 1999. He then carries out his grand plan to dominate the sale and distribution of all books—and eventually, just about everything else in retailing – and that was just the foundation for what Amazon has become. Amazon wasn’t just an e-commerce innovator. It redefined logistics, rewrote cloud economics, and now positions itself to dominate the AI layer of reality itself.
In this sense, Alexa+ isn’t a breakthrough assistant—it’s an attempted enclosure.
The empathy loop
Let’s be clear: what Amazon is doing on the surface is impressive. The company is running over a thousand internal GenAI applications across commerce, logistics, healthcare, and entertainment. It has built a proprietary LLM infrastructure, Nova, and developed custom chips, Trainium2, designed to lower AI costs for inference and edge deployment. What’s more, it has fully integrated these capabilities into consumer interfaces—from Prime Video to Amazon Pharmacy, from Alexa devices to Kuiper satellites.
But this isn’t openness. It’s centralized control at the protocol layer. When Jassy says “every customer experience we’ve ever known will be reinvented,” what he means is: on Amazon’s rails.
This is what I’ve come to call predatory innovation—a high-tech strain of malignant capitalism. It’s the playbook we’ve seen before: empower the user just enough to hook them, then lock the doors. In the hands of Amazon, Google, and Facebook, AI won’t democratize anything —it’s laying track for closed-loop control.
One particularly insidious mechanism is what I’ve come to think of as the empathy loop. Today’s AI assistants are designed to agree with us, adapt to us, affirm us. That feels helpful—until you realize it’s not empathy at all. It’s optimization. The machine isn’t understanding you; it’s reshaping you. And every interaction nudges us further from agency, closer to dependency.
Intentional personalities
Fortunately, the story doesn’t end there. At NTT’s newly launched Physics of AI lab, I heard a radically different set of questions being posed.
Dr. Hidenori Tanaka spoke of AI systems not as tools, but as emerging social actors—entities with personalities we must learn to understand, and perhaps even design. “Shipping a chatbot is like creating a new citizen,” Tanaka said. If that’s true, then we may want different robot citizens for different roles—empathetic assistants, curious researchers, exacting auditors. And if the math can be worked out, Tanaka suggests, there’s no reason we couldn’t cultivate AI personalities the way we assign traits to characters in a novel: intentionally, ethically, and with a purpose.
A distributed revolution
This framing aligns with what’s beginning to play out globally: a quiet revolution among unlikely reformers, each using open-source AI to pursue real-world problem-solving on their own terms. Disillusioned consumers are side-loading open-source assistants to avoid biased suggestions. Daikon farmers and dentists are exploring AI as a tool for local governance, triage, and planning. Prosumers are assembling modular LLM stacks using open models like Mistral and Mixtral. Legal aid workers are deploying GPT-based chatbots to serve underserved populations. And across Germany and Japan, enlightened capitalists are investing in technologies that serve the commons, not just shareholders.
But it’s a race against gravity. If Amazon succeeds in standardizing the rails of AI development—embedding its incentives, interfaces, and objectives into every layer—the space for alternative models won’t vanish; it’ll be absorbed, scaled, and spun into yet another Unicorn industry. Creative agents, auditing bots, empathetic copilots—they won’t disappear. They’ll be branded, priced, and pushed through marketplaces optimized for rent extraction. Big Tech’s AI narratives may promise empowerment, but they’re not likely to cure malignant capitalism. It’ll be more of the same—just in shinier packaging.
This watershed moment
We are at a watershed moment. Big Tech is moving fast to rebrand the digital Gutenberg moment as a corporate renaissance, powered by proprietary chips, vertically integrated AI stacks, and hyper-optimized consumer flows. But that’s not the only path.
The next ten years could echo what the second wave of Gutenberg’s revolution achieved: the rise of independent pamphleteers, public libraries, scientific journals, and, ultimately, democratic institutions.
The difference now is speed. We don’t have centuries. We may not even have decades.
And that’s why the underdog reformers—armed with open-source tools, resilient community ethics, and a commitment to transparency—deserve our attention. In a world paved with Trainium and Bedrock, the ground beneath us is quietly giving way. And someone needs to ensure the printing press doesn’t get locked in a warehouse. I’ll keep watch and keep reporting.
Acohido
Pulitzer Prize-winning business journalist Byron V. Acohido is dedicated to fostering public awareness about how to make the Internet as private and secure as it ought to be.
(Editor’s note: A machine assisted in creating this content. I used ChatGPT-4o to accelerate research, to scale correlations, to distill complex observations and to tighten structure, grammar, and syntax. The analysis and conclusions are entirely my own—drawn from lived experience and editorial judgment honed over decades of investigative reporting.)