The attention economy isn't innovation. It's a toll booth, and you're paying it with the only asset you can't buy back.
I want to say something that the industry I came from would rather I didn't.
The biggest technology companies in the world are not, primarily, technology companies. They are advertising companies wearing very expensive engineering. And once you see the revenue underneath the product, you stop asking why the product never really gets better for you. It was never built for you.
In 2025, Meta reported roughly $201 billion in revenue. About $196 billion of that came from advertising. That is nearly 98 cents on every dollar. Alphabet crossed $400 billion in annual revenue for the first time, and roughly three quarters of it came from Google advertising across Search, YouTube, and the Google Network. YouTube ads alone did over $40 billion.
Those are not side businesses. That is the business. Everything else is a delivery mechanism.
Advertising is not a strategy. It's a ceiling.
Here is the part that bothers me most as a builder.
When a company makes 98% of its money selling your attention, every product decision downstream is quietly constrained. You cannot ship the thing that gets a user in and out in ninety seconds if your revenue depends on keeping them for ninety minutes. The incentive is structural. It doesn't require anyone in a room to be a villain. It just requires a P&L.
So the innovation goes somewhere. It goes into targeting. Into ranking. Into ad formats, creative automation, bid optimization, impression volume, price per ad. Meta's own reporting credits AI ranking improvements with lifting ad clicks and conversions. That is genuinely impressive engineering. It's also engineering aimed at the advertiser, not at you.
We have spent two decades building the most sophisticated persuasion infrastructure in human history and we still cannot reliably answer a simple question: where should I eat on Thursday, and can I actually get in.
That gap is the story.
Watch the pivots. They always land in the same place.
Meta is the clearest case study, because they keep telling us they're building the future and then quietly rebuilding the toll booth.
First it was the metaverse. They renamed the entire company for it. Years and tens of billions of dollars went into a virtual world that most people never entered twice. Then it was AI, with capital expenditure guidance for 2026 running as high as $135 billion.
And then, in May 2026, came Meta One.
If you missed it, here is what the future looks like. Instagram Plus and Facebook Plus at $3.99 a month. WhatsApp Plus at $2.99. What do you get? Profile customization. Super reactions. Story viewer statistics. The ability to keep a Story up past 24 hours. Exclusive themes, custom ringtones, premium sticker packs. Above that sit Meta One AI tiers at $7.99 and $19.99.
We renamed a company after a new dimension of reality and arrived at paid emojis.
I am not being glib. I want you to sit with the actual scale of it. In Q1 2026, all of Meta's non advertising revenue combined, subscriptions plus hardware plus glasses plus headsets, came to about $1.29 billion. The ad business did north of $55 billion in the same three months.
So the diversification is real in the press release and rounding error on the income statement. Ninety eight percent of the company is still one thing. Every pivot so far has been a new surface to put an ad on.
Then TikTok raised the stakes
TikTok did something clever. It let users make the content, which meant the supply of things to look at became effectively infinite and nearly free. Brilliant business. Brutal experience.
Scroll your timeline honestly. Somewhere in there is something you wanted. Around it is a bunch of stuff you did not ask for, do not care about, and will not remember. That is the trade. You get one good thing for every twelve you have to sit through, and the twelve are the product.
And it isn't only ads anymore. TikTok Shop takes a referral fee on sellers, around 6% in the US on most categories, before you stack creator commissions, fulfillment, returns, and the ad spend sellers have to run just to be seen. Practitioners peg the real all in cost of selling on the platform at somewhere between a third and half of revenue.
So the platform monetizes your attention on the way in, and the merchant's margin on the way out. You are inside a store where everything is loud, everything is sponsored, and nothing is verified.
Noise as a business model.
And now the AI companies are doing it too
I had hoped this generation would be different. It isn't, at least not yet.
In 2024, Sam Altman stood on a stage at Harvard and said advertising was a last resort for OpenAI. He said combining ads and AI unsettled him. He said ads fundamentally misalign the user's incentives with the company serving them.
He was right. That is the most accurate diagnosis anyone in this industry has offered, and it came from the person who then went ahead and did it anyway.
On January 16, 2026, OpenAI announced ads in ChatGPT. Testing began February 9 for logged in free and Go tier users in the US. By July 22 they had opened a self serve ads manager where brands buy against conversational context. Independent tracking in 2026 found roughly three quarters of shopping related questions asked in ChatGPT now return a sponsored placement. Internal projections reported by Axios put ad revenue at $2.5 billion this year, and as high as $100 billion annually by 2030.
Read that again. The highest intent moment, the one where a real person is actively trying to decide something, is the first inventory that got sold.
And here is why it happened. It was not a product insight. It was a diversification failure. ChatGPT reportedly has hundreds of millions of weekly users and only a small single digit percentage of them pay. Compute costs are enormous. When the only lever you have is scale and the only asset you have is attention, you end up at advertising. Not because it's the best model, but because it's the available one.
Same trap. Newer building.
So I'll ask the question plainly. Where is the innovation in the technology industry? Does all of it, eventually, resolve back into advertising and the pursuit of your attention? Because from where I sit, the most brilliant minds and the largest balance sheets in the world keep arriving at the same answer to a question nobody asked.
The algorithm question
I love algorithms. That is not a contradiction. My whole thesis is built on them.
But you have to be precise about what an algorithm is optimized for, because that is the only thing it will ever actually do. Today's consumer algorithms optimize for dwell time, session length, return frequency, and impressions served. They are extraordinary at that.
They are not optimized to help you decide faster. Nobody is paid when you decide faster. Advertisers want eyeballs, watch time, and eventually conversion, but the whole funnel is built on a premise I no longer accept: that the way to sell you something good is to first waste an hour of your life.
The premise doesn't hold. It never held for luxury. Nobody's best experience with a great brand started with a pop up.
What we're building at LUXE AI
We built Luxi on the opposite incentive.
Luxi is an agentic concierge that sits on top of your wallet and your preferences and does the one thing the attention economy structurally cannot do: get you out of the app and into your life.
Give you the right recommendation, not the highest bidder. Book the table that is actually worth your Thursday, including the ones that are hard to get. Surface the perks, rewards, and benefits you already pay for and never use, because there is roughly $360 billion in value sitting unclaimed in cards and memberships people already hold. Verified, credible, current information, so the decision takes ninety seconds instead of nine tabs.
Our success metric is not how long you stayed. It's how fast you got what you wanted, and how good it was when you got there.
That is a different company. It has to be, because you cannot build that product on an attention P&L. The incentives would eat it within two quarters.
Time is the one luxury nobody has figured out how to manufacture. Everything else in this category is renewable. Your Thursday night is not.
Why I'm the one building it
I spent more than fifteen years inside luxury. Gucci. Tom Ford. Sotheby's. Modern Luxury Media. Rooms where the aesthetics were flawless and the data was immaculate.
And here is what I saw over and over: we knew an extraordinary amount about the customer, and almost none of it was ever handed back to her. It went up. To the advertiser, to the media plan, to the attribution model, to the brand. Never down, to the person who generated it, in a form that made her life measurably easier.
That asymmetry is the thing I could not stop thinking about.
Then I built a newsletter, Dose of LUXE, and grew it past 100,000 subscribers. Just knowing what people actually wanted and being right about it consistently. That taught me something the ad industry keeps trying to buy its way around: relevance beats reach, every time, and it is cheaper.
And here is the part I am proudest of. Open it. There are no ads running down the side. There never have been.
We built the partnerships differently. When a brand works with us, they don't buy an impression. They give our readers something. A perk. A table held back. A rate you cannot get anywhere else. Early access. Something real that only exists because you're on this list.
The perk does not buy placement. That distinction is the whole thing. A partner's profile, their article, their product, their room does not rank higher because they gave us something. It simply performs better, because the reader is being handed actual value instead of an interruption.
That is the entire model in one sentence. Do not sell the reader's attention to the brand. Make the brand give the reader something worth their attention. Everyone in that exchange comes out ahead, including the brand, which is the part the industry still refuses to believe.
Ten years as an entrepreneur, and the conviction has only sharpened. Whoever owns the consumer's portable preference profile, and returns the value of it to the consumer instead of extracting it, owns the next layer of this industry.
I am not building another place to scroll. I am building the thing that ends the scroll.
Luxi is what happens when the algorithm finally works for you.
Nyah is the Founder and CEO of LUXE Intelligence, Inc., the company behind LUXE AI Concierge and Ask Luxi, an agentic concierge for premium lifestyle discovery, dining, and rewards.
Sources: Meta Q4 and full year 2025 results and Q1 2026 reporting; Meta One subscription launch, May 27, 2026; Alphabet Q4 and fiscal year 2025 results; TikTok Shop Seller Center fee policy and 2026 seller cost analyses; Sam Altman, Harvard fireside chat, 2024; OpenAI ChatGPT advertising rollout reporting, January through July 2026; Axios reporting on OpenAI internal advertising revenue projections; OtterlyAI 2026 sponsored placement research.
