This is what veridive believes, and why it exists. Call it our manifesto.
The most valuable knowledge of our time is no longer being written — it's being spoken. The expert who would once have written a paper or a blog post now goes on a podcast, gives a talk, sits for a three-hour interview, or records an explainer. Founders explain their thinking on YouTube. Scientists go deep on podcasts. Investors, doctors, builders and operators say what they really think out loud. We call this vast, growing, mostly-unsearchable body of human knowledge the spoken web — and almost none of it is answerable today.
The scale is staggering
- Hundreds of hours of video are uploaded to YouTube every minute.
- By some estimates, YouTube alone carries on the order of billions of spoken words a day — the equivalent of many encyclopedias, daily.
- YouTube is effectively the second-largest search engine in the world, with billions of searches a day.
- Yet a tiny fraction of videos ever cross even 10,000 views — so the vast majority of what's said is, in practice, invisible.
An immense library of first-hand expertise exists — and we have almost no way to ask it a question.
Why it's broken today
- You can't read a video. You can't skim a four-hour conversation for the 30 seconds that actually answer you. Watching is linear; knowledge-seeking isn't.
- Search surfaces popularity, not answers. Roughly 70% of YouTube views come from the recommendation algorithm, not from search. Discovery is tuned for engagement, not for "who actually answered my question."
- The best answer is often buried. The clearest explanation of a hard idea might live in a 500-view video from a real practitioner — effectively undiscoverable next to a 5-million-view clip that never gets to the point.
- Text tools can't see inside. Google indexes the open text web. General chatbots don't systematically watch, read and cite video to the exact second. The spoken web falls through the cracks.
Why now
For most of the internet's life, making the spoken web answerable was simply too expensive — transcribing, reading and reasoning over billions of hours of audio was a fantasy. That changed. The cost of the AI needed to understand language has fallen by orders of magnitude in just a few years. What was economically impossible is now possible — and someone is going to build the answer layer for spoken knowledge. We intend for it to be veridive.
What we believe
- Knowledge should be answerable, not just watchable. You should be able to ask a question and get an answer — with the receipts.
- Truth needs a timestamp. An answer you can't verify isn't worth much. Every claim should trace back to the exact second it was said. See exact-second citations.
- The best source wins, not the biggest. Relevance and authority should beat view count and channel size. See how we rank.
- Credit the people who did the thinking. Every answer points back to the original creator. See how we help creators.
Our answer
Make the spoken web answerable: read what's actually said, judge sources by whether they truly answer the question, and return the specific moments — cited to the second. That's veridive — the answer engine for the spoken web.
Figures above are illustrative, drawn from public reporting on YouTube's scale, to convey the size of the opportunity.