Remix.run Logo
cmrdporcupine 2 days ago

This is just pedantic. "Algorithm" is obviously shorthand for: a recommendations system that shows me things I didn't explicitly opt into.

Compare e.g. Mastodon vs Twitter or Bluesky. The former simply won't show you anything you didn't explicitly subscribe to, and there's no hidden ranking system.

The law is not a computer program. It is up to human interpretation. The law merely needs to define the intent, which is actually fairly easy to explain: you're not a common carrier if you're mediating and promoting and ranking and pushing beyond what the user has subscribed to with their choices.

You can get technical that "sorting" and "filtering" is a form of that, but you'd be applying the lens of a software engineer, not a lawyer.

viktorcode 2 days ago | parent | next [-]

It is pedantic, and you have to be pedantic when talking about laws and regulations. The vaguely written laws have the tendency to be interpreted in the most restrictive way possible by the executive branch.

> "Algorithm" is obviously shorthand for: a recommendations system that shows me things I didn't explicitly opt into.

In that interpretation that is applicable to any form of broadcast, including TV and radio, driven by the user ratings of their previous programs.

cmrdporcupine 2 days ago | parent | next [-]

You haven't seen any proposed law. Just someone's choice of word that got your back up.

wyre 2 days ago | parent | prev [-]

This is hacker news, not a law board. use good faith.

AlienRobot 2 days ago | parent | prev [-]

I understand how you feel but you can't be like "I want the law to make things worse for these businesses" and then when asked to define the boundaries of the law you say "that's pedantic."

In many cases technology laws are myopic in that they only see the most massive websites and forget that there is a whole www outside facebook. Is sorted by likes/upvotes a recommendation? Is the total of likes of your friends a recommendation? Can only data points from the last week be considered? Can there be a falloff by age?

At which point does the weights of the variables start to constitute a "recommendation"?