SearchGPT represents a genuine inflection point for how information flows on the internet. Unlike Google, which returns a ranked list of links and lets users navigate site by site, OpenAI's prototype answers questions with synthesized text pulled from multiple sources, then credits each site with a link. This sounds like a minor difference, but it upends the entire attention economy that has funded digital media for thirty years.
The problem is obvious to anyone who understands how publishers survive: traffic drives value. When someone searches for concerts in Porto in August on Google, they visit multiple websites—a festival site, a ticketing platform, maybe a local newspaper. Each visit is a chance to display ads, collect data, or convert to a paid subscription. SearchGPT collapses that journey. It answers the question directly, usually with enough information that users never need to click through. The link is there, but it's ornamental now.
This has already played out once with Perplexity, a search engine that does exactly what SearchGPT promises. Perplexity aggressively crawls content without explicit permission and has faced multiple lawsuits as a result. OpenAI learned that lesson. They watched the New York Times litigation and understood that to scale this properly, they need permission—which means paying. The economics flip: instead of sites paying Google through SEO optimization to rank higher, media companies will now negotiate directly with search platforms for data access and visibility. Reuters, Reddit, The Atlantic—they are becoming suppliers to search engines, not just beneficiaries of traffic. Some deals are already happening. Reddit blocked all search engines except Google and negotiated exclusivity. The Atlantic sold OpenAI access to its news and opinion content. This is rational, even if it feels backwards: guaranteed revenue beats the diminishing traffic from an AI-mediated search result.
The second consequence is architectural. The internet's design assumes websites matter. Under this new model, a website becomes a content repository—valuable for its data, not necessarily for its visitors. The real interface becomes SearchGPT itself, or whatever comes next. Attention consolidates further into platforms that can aggregate and synthesize information. That's bad for fragmentation and good for centralization, though it may ultimately force media companies to invest harder in quality and exclusivity rather than SEO tricks.