Why Do Fast Pages Beat Official Company Filings?

Parag Agrawal says fast pages beat an official company filing because people will not dig a public company's headline revenue out of the authoritative source. He leaves the PDF page open, naming page 73 only as one possibility, and treats the filing with the SEC as the authoritative number.

If Google cared about authority and showed the filing first, the PDF takes three and a half seconds to load instead of the second another page needs. The reader is already frustrated, and finding the answer still means grabbing through to page 77. People are lazy and skip the work. Lots of people need the information, so a whole class of web content puts it on a fast page, above the fold. It is 99.99% right, so a visitor is not ready to treat it as 100% wrong. It adds real value by picking, out of a 300 page earnings report, the bits that belong above the fold. The labels for the result are pre-AI human slop, or catering to a lazy human and winning at SEO.

Agents are not made to click around, fumble around, and grab a PDF. The move is to take an excerpt from the most authoritative place on the web and try to bring it into the agent's context window. Agents are not forced into the trade-off humans lived with, because standard browsers and protocols never figured out how to point exactly and consistently, everywhere on the web, to the exact right highlighted tiny paragraph. Agents get the excerpt brought into context and then figure out what comes next.