Uploading became a privilege, something that normal average people don't do. Earlier people ripped their own CDs, DVDs, magazine scans, digitized their VHS, recorded shows with their TV card etc. and shared these files. I don't know anybody who has ever created a torrent themselves. Everybody just downloads and seeds the pre-packaged stuff from the centralized sites, much like the walled garden philosophy of app stores.
But your experiences sound like what goes on in the public trackers. It's reasonable that folks who go to Pirate Bay to get their TV shows are not going to be as engaged as those who go to TVV, MS, or MTV. There's also the security risks of letting untrusted users upload files that may get consumed by thousands of people in the first hour.
In private trackers, I have often found that uploading is encouraged, going so far as to provide very detailed guides in their wikis showing how to rip the media, prepare the torrent, and upload it to the tracker. If a better quality version can be uploaded to trump the previous version, that is also encouraged. Learning, quality, and appreciation are cornerstones of the private tracker ecosystem, in my experience.
For that, you want Direct Connect [0]. As a one sentence pitch: IRC for file sharing ;)
A few years ago I was very interested in IPFS but once I learned about the limitations it was pretty much not suitable for me. If you really insist on using IPFS then your best bet would be to have a central tracker host an IPNS based site and publish its database index as a series of json files. The javascript client will then have to query the index on its own in the browser.
The alternative would be to download the entire database and create a local index. In theory that's not a bad idea but over time your database will grow to several gigabytes. That's not comparable to just going to whatever site exist today and submitting a search term but it could be highly resilient.
The fundamental problem is that untrusted nodes can't provide search services. It's possible that they end up redirecting you to a fake listing. Blockchain style consensus doesn't work for something that requires responses immediately like a search engine.
Also bittorrent trackers are slightly more complicated than just a file store
The main reward for torrent uploaders is fame, would that translate well in IPFS?
Setting the precedent that media companies can seize domain names and force ISPs to block access to search engines if they don't like some of the search results (while ruining the lives of the people running those search engines) is arguably a high price to pay for this improved search service.