The elephant in the room is the ad-driven world we live. People need to be manipulated into constantly buying things in order to create an illusion of progress and doing work that matters. If facebook, spotify, youtube etc... did not exist, nothing of value would be lost. Even without the megacorps, the technology was already available for people to connect, create and share. In such a world, discoverability is not a problem. Communities would have been small and fragmented. There would be no winner takes all dynamic, sucking up the creative air from other people. The early decades of Internet held a promise of such a world. But then the smartphone came along and lead to market consolidation and created the discoverability problem.
> In such a world, discoverability is not a problem
oh, well... my two cents:
Before smartphones was the browser web. We had discoverability issues because very few search engines had a near monopoly (to put it politely), and there was no univerally good method to filter the magnitude of sites out there
Before that near monopoly we had a multitude of search engines, we even had engines-of-engines (look up "meta search"). Back then, we suffered from discoverability issues because of the sheer magnitude of web sites out there (to put it briefly)
Before those search engines we had a multitude of indexes, or lists. We even had lists-of-lists. Back then we suffered from discoverability probems because, well, even if we had only a minuscule fraction of the sites we had now, there was still no univerally good method to filter the magnitude of sites out there
And before that we had archives (essentially FTP servers), BBSes, and even physical paper based magazines. Even back then we had discoverability issues...
Go all the way back to physical libraries, and you had discoverability problems still. Paper scrolls? same. Mud bricks? same.
And way before writing, finding that wise person who was in the know about something would also be considered a "discoverability problem" methinks...
That reminds me: back in the day when we only had 3 TV channels, people were already lamenting that it was no longer possible to be a "Renaissance Man"; there was just too much to know so finite lifespan implied one had better accept keeping up with only a finite set of interests.
(when I was a child, I guessed the local library had more books than I could ever hope to read. Say 300k books at 2 books/day ~= 4 centuries)
I do like this article, but it seems like he's reaching a conclusion of aggregators and just calling them "taste makers".
Reddit is ultimately the solution, but it was ruined because of the corporate clown car needing everything to be advertiser data consumer friendly. What I mean is that by creating a subreddit, you are creating your own aggregator. If your content is edgy, niche, etc. then it's almost impossible to find if not outright banned. And god forbid your niche actually does get popular, because then suddenly it must change by virtue of forced moderation, what's allowed, outsider influence, etc. All in the name of ads and getting a piece.
Advertising needs more regulation. In the West's 80s and 90s, we did get on track to crackdowns, but this was focused on children and protecting them. Now YouTube/Google/Alphabet is somehow able to avoid that legislation because..... ?
In terms of media I think a good approach to this is, ironically, to skew consumption towards the old. Something that’s been held as valuable for generations, or even centuries, is probably worth your time.
Today you don’t have to seek out the new, it’ll find you. It makes sense then to default to the things proven to be good, at least when you find yourself actively looking for something to read, watch, or listen to.
I have had more than a few "smart" people reiterate that they like suggested contents and they want suggestions to get better, at the cost of privacy. So yeah I agree with the author here.
This is why Google ranks responses, then added the side box and tries to answer the question directly. Most people just want the answer and click on the first link. A negligible number ever move to the next page.
This is of course the opposite motivation to that of the people who write the pages. They (mostly) want you to go to their page.
But those authors aren’t Google’s target audience; they want to show an ad to their real audience.
I’d rather have a page of links but I know I’m an outlier so I don’t bother to get upset.
> Now, the traditional way to solve this has been through algorithms. We rely on them to find patterns in what you read or listened to or watched and then identify other things which you might like to read or listen to or watch.
That's an odd use of "traditional". Maybe I come from an older tradition, but my solution is to forget passive discovery. Instead of hoping someone else's algorithm will identify things for you, actively search for it yourself.
Before I start, I hate every algorithm designed to pair me with content. It feels either overly simple (YouTube/Amazon), or invasive and weird (Google).
But.
How can I possibly sift through the MOUNTAINS of garbage on the internet to find good stuff? Who has that time?
I also find that I often experience the modern version of "no one gets me" quite frequently, which is that none of the algorithms really get me. Maybe I am interested in too many disparate things. Probably the worst offender here is the YouTube algorithm. YouTube really really believes I'm someone who wants a never-ending stream of Joe Rogan, Ben Shapiro, Jordan Peterson, etc. I have no idea how YouTube became convinced of that, and why it won't give up on the idea. I don't click on the videos, When they pop up in my shorts I immediately swipe past them, I can't even figure out how the kinds of videos I do watch would be connected.
I would actually love to have a good algorithm that would scour all of the content of the web and recommend just the things I would like, but all the algorithms that exist seem bad at doing that.
"click your profile icon in the top-right of YouTube and choose Your Data In YouTube again. Scroll down to YouTube Watch History and click the field labeled On. There, you can choose the Turn off option" -- Another method is to set Auto-Delete to 'anything older than 3 months'.
It seems to me that it would be in YouTube's best interest to provide me with accurate recommendations. The more content they can get me to watch, the more ads I view, which is how they get paid. It would seem that more accurate recommendations would encourage more view time and more add engagement.
Could be; I believe they believe their recommendations are just good enough to get you (or at least their modal user) to watch the ads for which they get paid the most, and if they made more accurate recommendations they'd have to serve ads that didn't pay as well.
Maybe that's too cynical, and accurate recommendations are difficult. I've observed I can do much better by searching for content on my own, so I do.
I don't think the issue with finding something suitable is that there are too many options to choose from. The problem is that search engines hate you, the data is shit (unsearchable) anyway, and people spent the last century brainwashing themselves into accepting ads and marketing saturating every facet of life.
> sifting through twenty five types of tomato sauce
The examples here are strange. For the everyday products, you can't find out which one is the good one because they are no meaningful differences between most of them. The range of variants/colours/labels were cynically engineered to overwhelm you, occupy more display space, etc.
> Discovery trumps creativity in the era of abundance.
It always did -- before this 'era', I mean. e.g. You can't get your novel published without getting the publisher to read it first. It doesn't matter if its good if you can't get it discovered.
> We need some way of navigating this, because total available information’s not going to go back down.
Large AI models are the solution. We invented something that can more or less understand the impossible amount of content out there, and distill it in ways specific to each individual.
I do have a bit of experience with LLM locally, to test the limits of what it can do consistently. 16000 token context windows don't seem like enough to me to tackle this task. I don't know that it's possible to distill my musical tastes into a small enough number of tokens, for instance. Nevermind that you need humans to review the music, currently; unless there's magical song tagging models out there i haven't heard of.
there are minimally invasive mechanisms, like kongregate and that .io game site use, "people also played ..."; most aggregators suck, though. Netflix has always recommended stuff ("For you" or whatever) that i have no interest in, amazon recommends things that i would never buy - or have already bought, including from amazon. Pandora was awesome before everyone uploaded every song to youtube, now it's just a collection of playlists that play the same 20 songs over and over each time you start it.
I don't use and haven't used spotify or any other "music" service, because of pandora experiences and youtube. Also in my car i prefer to listen to old time radio which is easily discoverable* enough for at least a few years of content. Just copy to a USB stick and plug into the car. And my big issue with podcasts is i don't enjoy listening to most (nearly all) people just "talking", especially if they're chewing the scenery to make it more of a "captivating" experience.
I do miss mp3 streaming sites, though. My favorite one went dark during the pandemic, is still dark, and shoutcast dot com has sucked for discovery for at least half a decade if not longer. I think it's mostly popular in non-english speaking countries these days, at a glance.
* OTRR - old time radio researchers, there's otrrpedia and an online player available, the content itself lives on archive.org - i have about a TB of radioplays/content from there that i rotate through (only) in the car.
