3 ms·
> Don't like it? don't follow it How did you find it in the first place? A search? Without any kind of filtering (that's an algorithm that could be used to man
by falcolas 2y ago
> Don't like it? don't follow it
How did you find it in the first place? A search? Without any kind of filtering (that's an algorithm that could be used to manipulate people), all you'll see is pages and pages of SEO.
Opening up liability like this is a quagmire that's not going to do good things for the internet.
- BlueTemplar 2y agoYou're getting off topic here though ? «Had Nylah viewed a Blackout Challenge video through TikTok’s search function, rather than through her FYP, then TikTok may be viewed more like a repository of third-party content than an affirmative promoter of such content.»
- pixl97 2y ago>not going to do good things for the internet. Not sure if you've noticed, but the internet seemingly ran out of good things quite some time back.
- falcolas 2y agoIrrelevant and untrue. For example, just today there was a highly entertaining and interesting article about how to replace a tablet-based thermostat. And it was posted on the internet, and surfaced via an algorithm on Hacker News.
- rtkwe 2y agoThe question though is how do you do a useful search without having some kind of algorithmic answer to what you think the user will like. Explicit user or exact match strings are simple but if I search "cats" looking for cat videos how does that list get presented without being a curated list made by the company?
- seydor 2y agoretweets/sharing. that's how it used to be
- falcolas 2y agoHow did they find it? How did you see the tweet? How did the shared link show up in any of your pages? Also, lists of content (or blind links to random pages - web rings) have been a thing since well before Twitter or Dig.
- skydhash 2y agoHow do rumors and news propagates? And we still consider that the person sharing it with us is partially responsible (especially if it's fake).
- jimbob45 2y agoWithout any kind of filtering (that's an algorithm that could be used to manipulate people) Do you genuinely believe a judge is going to rule that a Boyer-Moore implementation is fundamentally biased? It seems likely that sticking with standard string matching will remain safe.