"I'm here to tell you a story."

George Hotz welcomed a packed room at SXSW yesterday for his panel "I Built a Better Self-Driving Car Than Tesla" with an intro as buzzy as the event's title. Hotz introduced his company, comma.ai, saying "We are a company! I did that whole song and dance," and then sprinted on into a whirlwind of a presentation.

At most presentations at SXSW, there's an air of skepticism, or at least, reasonable human detachment. But at this presentation, people had come to be impressed. And they were. I'm not saying Hotz is a huckster — he's done enough to prove that he's talented — but he worked that hotel conference room like a preacher at an old-time tent revival.

Hotz was dressed simply but deliberately for his presentation, wearing a flattering black v-neck and jeans. His old teen-nerd aesthetic has given way to something more like a Fashion Police-approved Steve Jobs. Over the course of 30 seconds of pacing and sweeping arm gestures, Hotz captured the room by running through the "sordid tale" of the last eight months of his life.

As he tells it, 2015 went like this: he was at Vicarious, a "very secretive company" working on AI in the East Bay Area. "I read everything," he says, "I read all their papers." Then he quit. Then he went to Thailand; then he went to Vietnam; then he thought about stuff. For two weeks. Then he met with Elon Musk, who asked him to make a vision solution for self-driving cars that would rival the MobileEye tech Tesla was already working with. Then Musk offered him $12 million (minus $1 million for every month it took Hotz to work on the task). Elon Musk has had much to say about the accuracy of this story in the past, but the people in the room at SXSW either didn't know that or didn't care. They laughed on cue, and whispered feverishly during most pauses.

The middle 30 minutes were devoted to an in-the-weeds programming lecture —€” Hotz ran through explanations of neural networks, universal function approximation, Kolmogorov Complexity, and the human eyeball. After a while, he may have sensed he was losing the crowd, so he launched back into his Musk story.

Musk called him on his birthday, Hotz said, and changed the terms of the deal. Musk called him on his birthday, he repeated. He wasn't having it. So, he said, "I'll build it myself." Hotz bought a brand new Acura. The guys at the Acura dealership were nice. They said "coooool, dude, self-driving cars!" Hotz built a system that could drive his car.

Then the room sat in silence and watched the seven-minute video Bloomberg made about Hotz three months ago.

The presentation was sprinkled with jabs at just about every power player in tech and transportation —€” MobileEye is "the most overvalued $17 billion company" Hotz had ever seen; General Motors is just afraid it's going to become Kodak; Uber plays by too many rules; Google and Tesla are going to get schooled by the technology that Hotz made in his garage.

"We're coming for you Google," he said, which was met by booming laughter.

Hotz said his self-driving tech would ship by the end of the year for "$1,000... maybe a little less than $1,000." He spoke in the manner of an infomercial salesman offering us a once-in-a-lifetime deal.

He concluded his presentation by explaining that the transportation sector employs more people in the US and worldwide than any other industry, and that his plan is to eliminate all of those jobs. Starting with "the cashiers, the secretaries, the salespeople, those people are easy to replace with AI," and moving up the chain until the only people left with jobs are programmers. And then... the singularity, of course.

The audience stood and cheered for Hotz for a full three minutes after he stepped away from the microphone, then hoarded around him for handshakes and selfies. If applause is any indication of affection (which, probably), they loved his suggestion that he would dismantle capitalism by gradually eliminating employment. Nobody asked if that plan would entail a rough in-between stage. I guess I probably should have, but I was busy searching the room for another skeptical pair of eyes.

One audience member stood up, gushed about the presentation, then asked if his 1998 Ford Windstar will be able to use Hotz's tech even though it "doesn't contain anything that resembles a computer any more than an Etch-a-Sketch resembles a computer." Hotz smirked. Mr. Windstar would have to upgrade to a Honda, 2012 or newer, at least for now.

One nervous attendee did ask if Hotz could possibly mean what he said, in regards to the shipping date and the price for comma.ai's self-driving unit. "When I make a claim, it happens!" he shouted. "How much? Cheap!"

Hotz had them, and he knew it. With the mood in that room, I wouldn't have been shocked if someone started dancing with a rattlesnake. Regardless of whether his tech works (which a substantial Bloomberg profile involving a test-drive largely affirms) there's no way to know for sure whether Hotz's product really will ship by the end of 2016.

SXSW is a sweltering, often stupidly boring brand orgy, where panels with intriguing names often turn out to be little more than distillations of Wikipedia pages on a given topic. Most of the big companies don't even come out to play on this ultra-humid, teen-infested playground. And it costs about $900 to be here. People leaving Hotz's talk were calling him a genius, possibly because it was true, and possibly just because it would be really exciting to have witnessed the early days of a really smart person solving one of the biggest challenges in tech.

When Hotz promised to hack the iPhone, he did it. And that was exciting both as a technical feat and as the story of a 17-year-old kid going up against the Goliath of Apple. But now that Hotz has much bigger fish to fry — promising to hack tech's entire power structure, capitalism, and ultimately, millions of people's jobs — it's disquieting that an audience would remain so unquestioningly on his side. Jailbreaking a phone is one thing. But Hotz seems hell-bent on breaking a whole sector of the American economy —€” if he does, there will be much more at stake.

Correction: A previous version of this article stated that Hotz was offered a $20 million Tesla contract. The actual figure was $12 million, and was misheard. We regret the error.