Hacker News new | ask | show | jobs
by toss1 1788 days ago
I don't see how (although I might not object to some values of your quote).

It seems a straightforward choice:

1) Go forward as a Carrier and delete ALL features with a hint of editing, promotion, or recommendation, i.e., a simple straight chronological feed of other members' posts/feed specifically selected by each user, and maybe a search function.

2) Go forward as a Publisher and select, edit, recommend, promote, annotate etc. as much as they want.

With Option-1 they can avoid all responsibility for content, merely doing takedowns on items for which they get notices. With Option-2, they have the same responsibility of any publisher for their content (e.g., newspapers are still responsible for the 'Letters to The Editor' that they choose to publish, and I'm sure edit out profanities, etc.)

I expect that users would actually strongly prefer Option-1, although actual "engagement" numbers could decline, it might actually be a better advert platform.

It also seems that viral content would still exist, but be more 'natural', i.e., not enhanced by algorithms, since you Alan could see something from Bob and repost it, which would be seen by Chris, who is subscribed to Alan but not Bob, and Chris could repost it to be seen by Debbie, who subscribes to neither Alan nor Bob...

So, I'm curious how you see that this would this kill/outlaw social media?

1 comments

I think option 2 is a non-starter. If you're going have newspaper-like liability, I don't think anyone can afford to do that.

That leaves option 1. I suppose such a thing could exist, but it would be very different from social media as it exists. It sounds like a mash-up of twitter (accounts, follows, retweets) and 4chan (minimal moderation). Which would be interesting.

But you're still talking about basically making anything resembling current social media sites illegal.

You're also probably making any niche forum illegal too, unless it's niche enough that the operator can reasonably subject every post to prepublication review to try to avoid liability.

For sure, the same rapid-promotion algorithms would not be workable at scale if they were to run at current speeds.

But would it be so bad to have a system that is primarily unmolested by algos (i.e., mostly Opt-1) but with a slower algo that promotes more judiciously might make for a much less toxic SocMed environment?