Hacker News new | past | comments | ask | show | jobs | submit login

I just don't agree with this sentiment. I don't work for twitter or any social media company, but it strikes me as their prerogative to ban content deemed unsafe if they don't have the means or wherewithal to properly police the content. From an engineering standpoint, how exactly do you propose to scan fiddles for objectionable content. With an image link, you could throw a neural net at it and at least tag it as nsfw (or scan a few images in a linked page).

And this isn't related at all to the bots and fake accounts (which I think is the bigger problem). But in the context of your argument, this is just non sequitur.




Ok, but they’re not blocking CodePen etc. And furthermore, you don’t even need a JSFiddle/CodePen whatever, you can run it on any website that you can edit code on! Is github.io next to be blocked? How about any unrecognized website? That JSFiddle has been targeted by this action is absurdity.


> their prerogative

Thus any business decision isn't a problem? Whether it's their prerogative, it still a problem in tech.

> to ban content deemed unsafe

Like any links? Or even text itself? The only thing that makes JSFiddle "worst" is how easy it is, but even then almost anything else is just as easy. If there's money to be made too, unless you block everything that cost less than the money to be made, what you do won't stop it.

Why not just put that warning over EVERY single links and not block anything? Do a white list instead.


> their prerogative

Thus any business decision isn't a problem?

Depends on your definition of "problem"... and thus your definition of who should care:

- legally (government): any business devising that comply with law are not a problem

- financially (investors): any decisions that increase profits are not a problem

- morally/ethically (users): any business decisions that you personally are okay with are not a problem.

Each of these has its own correction mechanism: persecution, lack of funding, customer outcry & abandonment.

In other words: there are ways to reverse bad decisions.

The only thing that makes JSFiddle "worst" is how easy it is

As pointed out in other comments, the problem is that anyone can anonymously create a malicious JSFiddle that runs undesired code. You could make a car that any website that similarly allows anonymous code execution should have that warning or be blocked. Most links, however, are better attributable. (Eg, require account creation).


> Most links, however, are better attributable.

I strongly disagree on that. The fact that you have an account behind it doesn't make it attributable at all. There's nearly no verification on 99% of the internet. Some studies consider that 9 to 15% of Twitter accounts are litteraly fake [1]. TwitterAudit believe that 40-60% of Twitter accounts are fakes.

They aren't attributable to anyone except a username, which is worthless.

> You could make a car that any website that similarly allows anonymous code execution should have that warning or be blocked.

It was never about code execution but what they call "wallet code" which is what I did in another comment [2].

> Most links

You can easily register a domain anonymously. Most links are fine sure, most possible domains, aren't, which is my point. Show a warning (which they do on URL Shortener) instead of blocking a domain altogether on ALL links and use a white list (which would include MOST used domains) instead.

[1] https://aaai.org/ocs/index.php/ICWSM/ICWSM17/paper/view/1558...

[2] https://news.ycombinator.com/item?id=20124667


It is a problem. YouTube throwing off content makers. Facebook banning campaigns. playstore banning Dev accounts. iOS store banning apps. Google search removing search results. Twitter banning people. In a lot these cases the businesses and people behind these are victim without ever doing something wrong. Bad or no communication is something they have to deal with. Yes, in a lot of cases there is a justified reason. But the good should not suffer from the bad. And that is what's happening right now. Same with this Twitter link ban.


>> From an engineering standpoint, how exactly do you propose to scan fiddles for objectionable content.

But if this was about getting people to click more ads, Twitter would be throwing enormous amounts of resources at it.

I mean, let's not pretend that this is an engineering problem.


Engineering and cost to me are intricately intertwined, so yes it is an engineering problem. If they could solve it without affecting the bottom line (or even better improving the bottom line), I'm sure they would.




Registration is open for Startup School 2019. Classes start July 22nd.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: