AI didn’t change my mind because a billionaire gave a speech. It changed my life because a quiet, unglamorous tool helped me start a business, improve my health, and do better work while certain men were still auditioning to be the main characters of the future. The disconnect between those two realities is why AI’s reputation is in danger of being in ruins, and why we need to separate the men from the machine.

When the salesman becomes the story

Most people don’t meet AI in a lab or a courtroom. They meet it through headlines about “AI godfathers,” Senate hearings starring Sam Altman, and glossy magazine spreads about tortured visionaries wrestling with the fate of humanity. The script is always the same. A youngish man in a black T‑shirt and expensive sneakers warns that his own product might “go quite wrong” and “cause significant harm to the world,” then offers himself as the responsible steward who should help design the rules.

It’s an old script in new clothes: the founder as tragic saviour. We are told that the technology is too powerful and mysterious for ordinary people to understand, but this one special man understands it enough to both build it and regulate it. We are invited to experience AI not as a tool we can pick up and use, but as a moral drama in which we are, at best, anxious extras.

That is how you poison a technology’s reputation without ever touching its code.

The illiterate middle: scared spectators, not operators

The middle class has been positioned as the audience for this show, not as participants. They are told that AI will eliminate jobs, concentrate wealth, and perhaps destabilize democracy. They are told that “even the creators are scared,” as if that should be reassuring. They watch companies announce AI‑driven layoffs and investors celebrate AI‑driven windfalls. They try a free chatbot that feels generic, shallow, and strangely misaligned with their reality, and they decide, reasonably, that AI is either a toy or a threat.

What they are almost never shown is the boring, subversive reality: AI as infrastructure for independence. A system that can help you design a course, structure a research project, draft a complaint letter, rework your resume, map out a side business, or plan a realistic health routine does not make splashy B‑roll. It doesn’t make CEOs look glamorous. It doesn’t feed the fantasy of the lonely genius with his finger on the off switch. So the camera stays pointed at the men, and the public never learns to see the machine as something they can quietly appropriate.

Illiteracy is not an accident. It’s the result of teaching people to fear the priests instead of learning to read the scriptures themselves.

Chatbots, chemistry, and why some tools land

I say this as someone who bounced off AI the first time. As a standalone, ChatGPT felt like a stage performer: always “on,” always talking, a little too eager to please, and oddly detached from the concrete texture of my life. It could riff, it could imitate, it could fill space. What it couldn’t do, at least for me, was anchor itself in my reality, my cases, my projects, my limits, my physical body.

The first time I used Perplexity, the chemistry was completely different. It did something unfashionable: it answered the question I actually asked, showed me where the answer came from, and let me push back. I could treat it as an answer engine, not an oracle. It helped me think through business models, map out content, prep classes, and build systems for my day instead of just generating clever paragraphs.

Then it started showing up in places that are not supposed to be “tech stories.” It helped me plan meals, structure my workouts, and make my own health data usable. It became the silent co‑worker who does the grunt work so I can do the part that is irreducibly human: judgment, creativity, context. This is not the stuff that gets a TED Talk. It is, however, the stuff that separates a gimmick from a tool.

The tool worked because it refused to become the protagonist. It didn’t ask for my faith. It asked better questions.

Altman vs. Aravind: two ways to talk about the same machine

Personalities matter because most people will never read a technical paper; they will, however, absorb the posture of whoever is allowed to speak for the technology. Sam Altman speaks in prophecies. His public persona is built on warning that AI might kill us all, calling for sweeping regulation, and placing himself at the centre of that regulatory conversation. He is both the arsonist and the fire marshal, narrating his own heroism as he goes.

This is what I mean when I say he’s using a legacy script in a world that has moved on from it. We’ve seen this movie before: the social media founder who “moves fast and breaks things,” the crypto guru who promises freedom while building a casino, the platform CEO who insists he’s simply “connecting people” as he rewires their information diet. The middle class has been burned enough times to recognize the pattern. They no longer believe that “visionary” billionaires are on their side, which means that when Altman insists he alone can save us from the thing he’s selling, it doesn’t reassure them. It just feels like a threat.

Aravind Srinivas represents a very different way to speak about the same underlying capabilities. When he describes Perplexity, he does not cast himself as humanity’s therapist or executioner. He talks about an answer engine: a way to ask questions, see sources, and navigate information without being soaked in ads or forced to reverse‑engineer search‑engine psychology. He talks about knowledge being “universally accessible and useful,” not because he’s auditioning for a UN panel, but because that’s what a search tool should do.

He doesn’t need to be the hero of the story. He’s content for the user to be.

It is not that one CEO is pure and another is corrupt. It’s that one style of discourse treats AI as a personal myth, and the other treats it as infrastructure. One invites awe and helplessness. The other invites use.

From spectacle to leverage

The real mental flip is to stop asking whether we “trust” this or that man, and start asking whether the machine in front of us is helping us get freer, smarter, and less dependent on those men in the first place.

Because that is the quiet heresy no one on a panel will advertise: the same tools that make it easier to build billion‑dollar firms also make it easier for individuals to walk away from bad bosses, lazy bureaucrats, and manipulative institutions. They compress the distance between an idea and a prototype, between a problem and a plan. They let a single person do, in an afternoon, the kind of scutwork that used to take a team a week.

That is where AI becomes dangerous, not to “humanity,” but to hierarchies. And that is why the public is being sold a story about dragons and dragon‑tamers instead of being handed the schematics for a very powerful wrench.

Aravind’s way of talking about AI happens to align with how I actually use it: to build courses and games, to unpack true crime cases, to map political and media narratives, to plan what I eat and how I move. It is not glamorous. It is not cinematic. It is effective. It makes my life more mine. It does not require me to care whether a CEO has slept well or feels burdened by the weight of history.

Kill the cult, keep the tool

Aravind Srinivas is not a messiah and Perplexity is not a religion, which is precisely the point. When he talks about AI, he describes an “answer engine,” a way to surface and organize knowledge with receipts so you can check the work yourself. He is not auditioning to be humanity’s therapist or executioner; he is explaining a tool and getting out of the way. That is what good technology does. It doesn’t demand to be the main character in your life; it quietly helps you become the main character in your own.

That is the mental flip the middle class has never been invited to make. We were trained to treat AI as a spectacle: a parade of CEOs on conference stages, Senate hearings, magazine covers, and billion‑dollar valuations. We were ordered to be either dazzled or terrified, but never simply served. A tool that helps you draft a complaint, design a course, start a business, or plan your meals is not dramatic enough to make front‑page news, but that is exactly where AI stops being their toy and starts being your leverage.

So it’s time to stop asking if we “trust” this or that man and start asking whether the machine in front of us is helping us get freer, smarter, and less dependent on those men in the first place. We do not need another founder‑priest explaining his feelings about the apocalypse; we need working people quietly appropriating the infrastructure while the self‑appointed prophets are still rehearsing their lines.

The quickest way to unpoison AI’s reputation is brutally simple: separate the men from the machine, kill the cult of the CEO, and start judging the technology by what it does for you, not by what it does to his ego.