Jensen Huang has told The Verge there's a "0% chance" AI destroys jobs on the scale people fear. He's the CEO of Nvidia, a company whose entire market cap now rests on the world believing AI is about to remake everything. So take that number with the same pinch of salt you'd take from a man selling shovels telling you there's definitely gold in these hills.
I don't say that to be cynical for the sake of it. I say it because I've been doing this since 1986, and I've watched this exact pattern play out with mainframes, the internet, mobile, cloud, and blockchain. The person with the biggest financial stake in the hype cycle is never the person to ask "is this overblown?" It's like asking your estate agent if now's a good time to buy.
The founder-overclaim problem
Every big infrastructure shift produces two kinds of noise: the doom merchants who say it'll wipe out civilisation, and the boosters who say it'll fix everything and hurt nobody. Both are usually wrong, and both are usually talking their book — the doom merchants often want regulation that locks in incumbents, the boosters want you to keep buying GPUs.
Huang's specific claim — that AI won't cause mass unemployment — isn't obviously false. But "0% chance" is the language of a man who needs you to keep spending, not a man doing careful labour economics. Real technologists hedge. They say "it's complicated" and "depends on the sector" and "we don't fully know yet." Certainty at that level, in either direction, is a marketing statement wearing an engineer's coat.
What I've actually seen change
I run a small suite of content tools — RSSMasher, MarketMasher, BookMasher, Article2Video, VidMasher. Every one of them automates work that, ten years ago, took a human several hours: writing, editing, formatting, publishing. That work hasn't vanished from the economy. It's shifted. The people who used to spend four hours writing a first draft now spend forty minutes reviewing and shaping one. Some of those people got faster and more valuable. Some got made redundant because their employer decided one editor could now do the work of three.
Both of those things are true at once. That's the actual texture of technological change — not zero disruption, not apocalypse, but redistribution, often unevenly, often unfairly, and always slower than the headlines suggest and faster than the reassurances suggest.
What builders should actually do with this
If you're running a business or building tools, the Huang headline and the doom headline are both noise you can mostly ignore. The useful question isn't "will AI destroy jobs" — it's "which parts of my own workflow are now cheap enough to automate that I'd be daft not to, and which parts still need a human doing genuinely human work."
That second question is the one worth your attention. In my world that's things like: automating the grind of pulling together source content, generating first drafts, and handling the repetitive publishing plumbing — while keeping a human hand on judgement, tone, and anything that touches trust with an audience. The tools I build are deliberately aimed at the first category. I've never pretended they replace the second.
The tell is always the certainty
Whenever you hear a founder state a probability with total confidence about something this messy — 0%, 100%, doesn't matter which end — treat it as a data point about their incentives, not about the future. Huang has every reason to want you calm and spending. The AI-doom crowd often has their own reasons to want you frightened and legislating in their favour. Neither group has a crystal ball; they have a position to defend.
Forty years in, the only pattern I trust is this: the technology arrives, it's messier and slower than the hype, it's more disruptive and more permanent than the reassurance, and the people who do best are the ones who stopped arguing about the headline and got on with figuring out what it actually changes for their own work. That's the only bit of alchemy worth doing here — turning the noise into something useful, rather than picking a side in someone else's argument.
— Wayne