We're all missing the most important thing happening in media right now.
In the 1870s, Andrew Carnegie sent a letter to one of his employees about a new way of using a raw material crucial for their railroads business. It contained an order:
“Advertisements in every journal devoted to the mechanical engineering and architectural arts. Someone should read a paper upon it before the engineers. Your Franklin Institute should know about it. Indeed, I should within ten days have so advertised it as to render it almost impossible that any man of intelligence should be ignorant of the fact that an important step has been taken by us to advance the industry.”
Ben Wilson has a twist on this built from studying men like Carnegie, Napoleon, Rockefeller, Jobs, and more — he said: the shameless will inherit the earth. Carnegie. Jobs. Trump. The people who do a little too much and don't care that they went a little too far. Because when it comes to publicity, you don't want to make it possible for people to hear about you. You want to make it impossible for them to be ignorant of you.
Now read that letter one more time, and tell me how it's different from what a16z announced last November: “Timeline takeover. We offer 'win the internet for a day' as a service.” So what did change since Carnegie's days? The gatekeepers. That's it. As a16z explained it: “The old guard of legacy capital, legacy media, and legacy distribution can now be sidestepped; anyone can go viral, and gain relevance for a day.”
Using Naval's words, “you want to work in a world of permissionless capital, permissionless media, and permissionless product building.” But other than the gatekeepers deciding which story to publish, which name they want to raise and which to fall, most didn't change. It didn't. Carnegie was running timeline takeover in 1874 — the timeline was just made of trade journals and papers. So here's the first thing to understand about New Media: it's not new.
a16z states the problem better than most creators do: “Anyone can go viral, and gain relevance for a day. But what do you do with this?”
That's the right question — and notice it's coming from a VC, not from MrBeast. VCs are in the business of capital allocation, not attention. When capital allocators start building media pit crews, it means they've concluded attention is where the returns are. Their answer, though, is for their portfolio: control the internet for a day, delivered by their staff, like Carnegie's ads delivered by his staff.
If you're a creator, the answer to “what do you do with this?” has nothing to do with any single day on the timeline. It comes down to the only thing that matters in your world right now: where your atomic unit of value is being created, who controls it, and how it gets monetized at scale. Paul Graham said it simpler — “make something people want” — but in media, you need to make something people want AND to know precisely where the wanting lives. And to understand that, we need to understand Elon Musk's Idiot Index.
Elon Musk's Idiot Index is the ratio between what a finished part costs and what its raw materials cost. A high spread means the process transforming the raw material to the finished part is expensive — that's where the money hides.
Now apply it to New Media. The raw material is a recorded conversation, a written piece. The finished goods: clips, posts, newsletters. Two years ago the spread between those was enormous — human editors, days of labor. That spread is what every content tool on the market still monetizes today. But AI is compressing it to near zero, for everyone, simultaneously. In manufacturing, when your process innovation is available to all competitors at the same API price, it isn't an advantage — it's the new floor.
And that new reality is creating an uncomfortable truth: a business that sells reduction of production cost, in a world where production cost is going to zero, is selling a depreciating asset. “We make content cheap to process” is a 2024 thesis with a 2027 expiry. And a creator spending their time chasing more output — more stuff, more places, more of everything — is buying that depreciating asset with their time, attention, and money.
When manufacturing cost goes to zero, value migrates to the two things that can't be manufactured: attention allocation and trust.
Attention Allocation — deciding what deserves attention. Platforms, algorithms, taste-makers. The algorithm's asset is match data: knowing, better than anyone, which story redeems which attention-moment. That data compounds mechanically, which is why allocation scales like nothing else — and why, in 2026, it's off the table for 99.9% of us. The platforms are the matchmakers between stories and attention, clearing that market billions of times a day. You will not out-allocate the feed. I'd actually argue that this value, if I'd have to bet on anything, will be decided by your agent very soon. But for now we're still scrolling.
Trust — the parasocial asset that makes a specific human's story worth more than an identical AI-generated one. Creators will own this. Nobody else. And trust has an atomic unit: a kept promise. Every piece of content is an implicit promise — “an hour with me will be worth it” — and trust is nothing but the compounded record of promises kept. Accumulated one fulfilled expectation at a time; non-transferable, non-manufacturable, and destroyed faster than it's built. Its economic expression: the price premium a story commands because this specific person tells it. Ask any Founders Podcast listener — David Senra's asset isn't his episodes, it's a decade of being right about what deserves your hour.
Everything in between these two is commodity. Every creator and every business in this space must convert itself into an allocation business or a trust business — or get margin-compressed to zero.
This isn't a prediction. It's a rerun. All of media history is one repeated event: the cost of manufacturing claims on attention drops, and value migrates upstream. The printing press made text cheap — value moved to publishers (allocators). Broadcast made audio and video cheap — value moved to networks (allocators). The internet made distribution cheap — value moved to feeds (algorithmic allocators) and to individual creators (trust holders), gutting the middle. AI is the terminal event in this sequence: it makes the claim itself — the finished artifact — free. For the first time, the market clears purely on the real assets: who decides what deserves attention, and who is trusted enough to hold it.
So if you're a creator, the strategy writes itself: allocation belongs to the feed, the middle is commodity, and the only asset left standing is yours. Protect the promise. Own the trust.
Except there's a trap. And that trap is you and your finite time.
Carnegie demanded shamelessness, but he didn't place the ads. An employee did. Payroll did. a16z's founders get shamelessness delivered by a pit crew. The creator gets the sermon — post everywhere, post daily, be impossible to ignore — without the staff.
A creator has something worth hearing: the raw talent telling a personal story. But the world has finite attention, seeking things worth attending to. Everything between those two facts — editing, clipping, posting, timing, matching to audience — is friction. Friction is the only thing standing between a good idea and the attention it deserves. And right now that friction is paid in human hours: your hours, the same ones the promise is made from. Yes, you're using AI tools, but you probably spend over 20 hours a week editing, clipping, and posting. Every hour spent running the commodity layer is an hour taken from the only asset that survives it.
That's the bind that has quietly turned “creator” into a manufacturing job. You hold the one asset AI can't touch, and you spend your days doing the one job AI already does.
Look at the shape of that second job — editing video, writing copy, scheduling, distributing, analyzing. Bounded, repeatable, judgment-plus-taste tasks. That is exactly the shape of work agents do now. The commodity layer isn't just cheap to automate. It's supposed to be automated — it was never the part only you could do.
So here's the promise, and the whole strategy in two sentences: You keep the promise to your audience. Let AI handle everything past your story.
That's why I'm building Lucidream. Not another tool selling cheaper clips — cheap clips are the floor. An agent that runs everything downstream of the conversation: cuts it, captions it, publishes it everywhere, watches what works, and gets smarter about your audience with every post — so the story stays true to the promise, and the promise-keeper gets their hours back.
For 150 years, the shameless inherited the earth because they could afford the machine — Carnegie's payroll, Jobs' agencies, a16z's pit crew. Trust still can't be manufactured. Shamelessness, as of now, can be.
Go a little too far. Be everywhere, all at once. The people who do a little too much and don't care that they went a little too far. Because when it comes to publicity, you don't want to make it possible for people to hear about you. You want to make it impossible for them to be ignorant of you.
And that's my promise to you. I'm building Lucid so it's impossible for them to be ignorant of you, while you can focus on your own promise to your audience.