Ethan

Creating a paid or ad-supported client app for a website isn’t profiting off of content, it’s profiting off of the user’s desire for a better mobile experience. There’s no ‘stealing’, the developer never has access to nor purports to own any of the content themselves- it’s simply a voluntary intermediary for a user to access their own account with their own content feed.

That said, any client apps that run ads are dumb and will fail miserably. It’s awful for UX. Just so long as client apps can be monetized in other ways I think it’s fine to adopt a license that prohibits specifically ads.

This same story was posted yesterday, so I’ll rewrite what I did back then:

Most of this report is patently ridiculous. HRW asked people who follow the HRW social media accounts to please send in perceived instances of censorship they’ve seen for the Palestinian conflict social media, they got about a thousand examples from a self-selecting population, then published a big exposé about it.

There’s no comparative analysis (either quantitative nor qualitative) to whether similar censorship happened for other topics discussed, other viewpoints discussed, or at other times in the past.They allege, for example, that pro-Palestinian posters didn’t have an option to request a review of the takedown. The obvious next step is to contextualize such a claim- is that standard policy? Does it happen when discussing other topics? Is it a bug? How often does it happen? But they don’t seem to want to look into it further, they just allude to some sense of nebulous wrongdoing then move on to the next assertion. Rinse and repeat.

The one part of the report actually grounded in reality (and a discussion that should be had) is how to handle content that runs afoul of standards against positive or neutral portrayal of terrorist organizations, especially concerning those with political wings like the Hamas. It’s an interesting challenge on where to draw the line on what to allow- but blindly presenting a thousand taken down posts like it’s concrete evidence of a global conspiracy isn’t at all productive to that discussion.

Most of this entire report is patently ridiculous. They asked people who follow HRW’s social media to please send them instances of censorship on social media, get about 1,500 random examples from a self-selecting population, then publish a big expose about it.

There’s no intensive comparative analysis (statistical or otherwise) to other topics discussed, other viewpoints discussed, or at other times in the past. They allege, for example, that some people didn’t have an option to request a review of the takedown- is that standard policy? Does it happen in other cases? Is it a bug? They don’t seem to want to look into it further, they just allude to some sense of nebulous wrongdoing then move on to the next assertion. Rinse and repeat.

The one part of the report actually grounded in reality (and a discussion that should be had) is how to handle content that runs afoul of standards against positive portrayal of terrorist organizations with political wings like the PFLP and Hamas. It’s an interesting challenge on where to draw the line on what to allow- but cherry picking a couple thousand taken down posts doesn’t make that discussion any more productive in any way.

Honestly, BlueSky’s AT Protocol fixes pretty much all of these issues (save for having a single actor controlling things as for the moment it’s still in active development and not adopted by any other project).

Even if you never intend to sign up for or use their protocol, I’d give it a read- it’s a really fascinating system design:

https://atproto.com/guides/overview

I think you’re fundamentally misunderstanding how data is handled in federated systems. When an account from Threads interacts with your post or you interact with a Threads post, information is exchanged exclusively through Actions sent between the servers- never to or from your or their client and another server. It looks like this:

Client <=API=> Instance <=Action=> Instance <=API=> Client

They don’t get any information that isn’t already available publicly to any random user on your instance- no IP address or anything otherwise. Threads’ mobile app data collection has no bearing on their ability to collect information on you.

Edit: To be clear- there is theoretically a set of protocols in the ActivityPub spec that allows for direct client to server communication (unimaginatively called ActivityPub Client-to-Server), but it hasn’t been adopted by any current Fediverse software implementation that I’m aware of.

Threads’ daily app downloads have universally been in the range of 350k-700k for the past month. Mastodon’s MAU for the same time period has been 1.1 million.

More people downloaded the threads mobile app in the past three days then have interacted with Mastodon in any capacity for the past month.

Edit: Even your cited source pegs Threads’ lowest recorded daily active users at nearly half Mastodon’s monthly active users, and that was from before the app was made available in India and the EU

https://techcrunch.com/2023/12/04/threads-downloads-return-to-growth-as-x-adds-walmart-to-its-advertiser-exodus/amp/

Mastodon was immediately dwarfed from the very first day Threads was launched. Total Fediverse MAU has been hovering a but under 2 million, Threads first day user signups totaled more then 30 million. Threads’ growth has leveled off now but it’s still orders of magnitude more massive.

They wouldn’t be able to get it even from a federated instance as long as you don’t use their frontend.

I don’t know of any major instances that have enabled any of those… And all getting around it would take is to create an account on the instance- which for instances without admin approval can be done fully programmatically anyway so it wouldn’t even require human intervention, just a few extra lines of code.

They could already get all of your profile and post info… I could get that right now through the free api for every account on this and every other thread with a couple dozen lines of code.

Edit: I’m also unclear on how they would ever get your IP- if you never use their frontends the only IP they’d have access to is that of the server your account is hosted on… Which would only be your own IP for the extreme minority who host their own instance from their personal internet connection, and Meta wouldn’t be able to tell that that’s the case anyway.

“Consumer Reports factors build quality issues that require repair into our reliability calculations, but they are not weighted as heavily as more serious problems, such as those with the engine, transmission, or drivetrain.”

I’ll take a guess that this is the source of the weird counterintuitive results. EVs have a bunch of minor QC issues resulting from their comparatively newer and less tested assembly lines, while ICE vehicles have fewer but much more severe issues. Thus they’re rated as more reliable by CR even though the average cost of maintenance is far higher.

I’d be quite curious to know exactly what their weights and standards are for determining repair severity, because their current results don’t line up with any previous direct cost of maintenance studies.