On Friday night, the screens at Madison Square Garden lit up purple with "Just Married," and every brand with a marketing budget tried to shove its way into Taylor and Travis's wedding photos. Meanwhile, Target quietly remembered it owns a marketplace, and Shopify finally admitted out loud what developers have screamed for eighteen months: the app store reviews are fake. Three stories, one theme, everybody's chasing trust they didn't earn.

Taylor Swift married Travis Kelce, and every brand with a marketing budget tried to climb into the limo. The Knot rented four billboard trucks to circle Madison Square Garden. An antique shop posted a fake photo of its furniture being wheeled into the venue and pulled 11k likes for the trouble. Dior, Louboutin, and Cartier got their placements, and the earned-media math looks great on a slide: Dior around $15M, Cartier close to $6M.

Here is what nobody showed me. A single sale. This was a hype event, not a conversion event. 748 million TikTok views is a lot of noise, and noise is not a checkout button. If you are Nike or Sephora and you do not naturally belong at that wedding, squeezing your way in just looks like the guy who shows up uninvited and won't stop asking, how about the Knicks. Newsjacking has a shelf life, people. The first AI-slop furniture video is clever. The tenth is landfill.

Target Plus grew 60% year over year, which sounds heroic until you remember it is off a tiny base. Target has roughly 2M SKUs and 1,500 brands. Walmart carries 420M SKUs from 200,000 sellers. Amazon is playing a different sport, 600M SKUs and 1.9M active sellers.

What I actually like here is the discipline. Target is pushing the heavy, slow-moving stuff, the TVs and printers and laptops, off the store shelves and onto the marketplace, and keeping physical real estate for products that actually turns. Anyone who has stared at an empty Target shelf knows why that matters. Now show me the QR codes in the electronics aisle. Show me a back-to-school tie-in. It is July. Prime time. I am watching.

Eighteen months of developers screaming into the forums, and Shopify has concluded that the app store reviews are, in fact, gamed. Congratulations on the five-year-late memo. The new policy says you cannot ask for a five-star review, only a review "in a neutral tone." How do you enforce a neutral tone? AI sentiment analysis, reading your emails, obviously. Your outreach is getting scanned.

The bigger question is whether app stores even matter now. Nobody I trust picks software based on a star rating. I trust app store reviews about as much as the Lululemon dupe with 80k reviews that launched last Tuesday. Merchants ask peers. Buyers ask an LLM. And 80% of any app store is technically-valid slop you would never let a real client touch.

OpenAI launched Atlas in October, and by July, it was gone. Fidji Simo called it a side quest, and she was right to.

Here is what people keep missing. The browser is a tab. It is not a destination. Nobody wakes up wanting a new place to type a URL. They want the thing on the other side of the URL to happen without them lifting a finger. That is exactly why OpenAI is folding Atlas back into a Chrome extension, a beefed-up desktop app, and a cloud browser that runs errands while you go do something better with your afternoon.

I have said for two years that the real fight in commerce is the checkout, not the search box. Perplexity's Comet, The Browser Company's Dia, Google's Gemini Side Panel. Everyone is circling the same question. Who owns the moment the agent actually buys?

So no, Atlas dying is not a retreat. It is OpenAI admitting the browser was the wrong container for the bet. The bet itself, agents that transact on your behalf, only gets bigger from here. Watch where the checkout lands.