October 26, 2023

s16e17: Rude; A Financial Services Social Network?; We Should Improve Attribution and Sourcing Somewhat

0.0 Context Setting

Thursday, 26 October, 2023 in Portland, Oregon, U.S., where there's no way to prevent the events of last night in Maine, in the only nation where this regularly happens².

Man, gun amnesties/buy-backs work. Try them more, like in New York¹. Fuck it, at this point just reward people who turn in their guns with a year of free healthcare. Or no taxes for a year.

---

0.1 Hallway Track News

Hallway Track 003 You Deserve a Union

As of writing (~10am PT), there are still a few places left for today's Hallway Track, You Deserve a Union, at 2:30pm Pacific, 5:30pm East Coast, 10:30pm London.

Joining me on this Hallway Track are:

• Jacky Alciné, Union Organizer, formerly Code for America, and backend software engineer
• Ethan Marcotte, designer and author, You Deserve a Tech Union (code HALLWAY15 for 15% off at A Book Apart)
• Clarissa Redwine, tech worker, formerly of Kickstarter United, labor organizer and activist, member of the team behind Collective Action in Tech, and an organizer of the labor conference for tech workers
• Johanna Weststar, Associate Professor of Labour and Employment Relations at Western University, Vice-President of her faculty union, and co-author of Not all Fun and Games: Videogame Labour, Project-based Workplaces, and the New Citizenship at Work (code FUN2024 for 20% off, U.S. pre-order, Canada pre-order)

Register for Hallway Track 003 You Deserve a Union!

Next, on Hallway Track

Hallway Track 00 title tba, with Joanne McNeil and Tim Maughan on Thursday 2 November at 9am Pacific, 12pm Eastern, 5pm London, registration opening tomorrow.

Hallway Track 005 The New Luddites Seizing the Means of Computation, with Cory Doctorow and Brian Merchant onThursday 9 November 2023, 3pm Pacific, 6pm Eastern, 11pm London, registration opening Friday 3 November.

---

1.0 Some Things That Caught My Attention

Just a few small things that caught my attention this morning.

1.1 Rude

More software should let you be rude. Or irritated. You know, all the fluffy stuff like validate your emotions, it's totally fine for you to be angry or upset. All this bland interface shit, the surprise and delight feels, to a Brit, that kind of fake-American have-a-nice-day, either of the Californian/west coast kind smiling, or of the Midwest passive aggressive I want to fucking murder you while I smile.

I guess one of the analogies would be New York Software.

This came about when I was unsubscribing to one particularly annoying mailing list (which I should've unsubscribed to ages ago, so I've grown to resent it) and I just wanted to reply fuck off to the email. My mail client should be smart enough to interpret that as an unsubscription request and process it accordingly; macOS Mail can do this for certain mailing lists/mailing list software.

At the very least, this has reminded me that it's in my power to rename my spam training folder as "off you fuck!"

I want rude software. I want vindictive, nasty surprise and delight. Not that this is healthy, I suppose, but that at least having an outlet for it would be healthy. Like, I don't know, if there were a Beat Saber version of an email client where you could whack the shit out of spam in your inbox and get a workout, too.

Carrot Weather, on Apple platforms, is the only software I'm really aware of that is, at the very least, surly.

1.2 A Financial Services Social Network?

Elon "wants" the social network formerly known as Twitter to become some sort of financial services be-all/end-all app and who knows if he's going to be able to yell at enough people to have his dream of a financial services app for a few million batshit raving fascists.

But!

Struck me, this morning, that Venmo is already a pay-for social network, at least with its public feed.

(I do not understand why people leave public posting on in Venmo, other than not knowing it's the default and not knowing/bothering how to turn it off. Anyway, internet friend Jared Spool was interviewed recently on the power of defaults⁵.)

Seriously: pay $1 to your mates or whatever and you get to post not only on their feed, but also the public feed! All Venmo would have to do would be following relationships so you can... see what people post, er, pay for things?

1.3 Public Means Public

Confluence of a couple conversations online with Mark Blacknell⁴ and Tindra³ about what Public means.

The first was a mismatch of systems. Given the purpose of a system is what it does, likes aren't public on Bluesky. But likes are public on the Bluesky API. So Tindra pointed out that someone got upset that their likes had been made public. But again, the apparent what-it-does of the Bluesky system from the vast majority of their users' point of view is to not expose likes, whereas the purpose, the what-it-does of its API is much wider.

Should one reflect the other? APIs can (and do?) reflect functionality that isn't necessarily exposed in any front-end, so fine. At least, public APIs do! Should retrieving likes have been made available in the API? I'd argue yes.

The second was Blacknell's point on the Venmo public feed and public feeds in general, that "almost everyone with public feeds have no idea they're public".

This makes sense to me. Until you've experienced it -- i.e. you're the Main Character -- truly "public" is abstract. Until the nature of what you've made public has been used in I suppose a wholly unanticipated way, "public" is abstract. I am relatively certain that those involved in the January 6 riots in the U.S. didn't anticipate, at the time, how their public social network posts would be used. I'm excluding here the kind of people who would bandy around words like opsec, in many cases just cosplaying.

There's public-because-it's-harmless, and then there's horrifying public, again because of context, I think. (And framing!) What would it be like if you knew that what you were sharing on Twitter was visible to over 100 million people?

There is a tradeoff here, I suppose. You can't share to a small group. You are in practice sharing to a small group (those following you), but that barrier is porous because on one end, you can be screenshotted and at that point you're done. But the system-is-what-it-does builds porosity in through retweets and quote tweets or whatever the mechanism is on that platform. You know about that porosity, you get notified about it when you get retweeted. Would it be different if every time you got retweeted or your post had another chance of being on a/someone's feed, you were told how much larger your total audience was?

One of the first things people get taught, or realize, once they've become Main Character is the value or necessity of going private. But, last I checked, you'd need to be told you'd gone Main Character by someone, or by the time you know, it's too late. What would it be like to set a threshold? In a social network that prioritized, say, porosity-to-a-limit, which would mean empowering people to define and manage their boundaries (and therefore their privacy), then why not let someone say "as soon as I cross this threshold of visibility, then automatically make my post or my profile private"? This won't prevent the screenshot problem, but saying it's not worth doing because people can screenshot is disingenuous, it's like saying gun control in the U.S. is useless because people could still get guns. But still some control is better than no control, no?

1.4 We Should Improve Attribution and Sourcing Somewhat

So here's a dumb idea, no idea how workable it would be.

There's the joke that most social network posts are just screenshots, more specifically, screenshots of tumblr posts (or reddit posts, I guess).

Is this problematic? It is in, say, moderation of violent content. Just like with CSAM, the large social network providers maintain hashes of prohibited images and video (whether prohibited by their platform, or legally). Yes, these hash databases are not complete, fine.

But! If you're worried about attribution, then:

• hash all images
• for each hash
  • perform a reverse image search
  • perform text recognition and a search on recognized text
  • "do magic" to determine likely source URLs
  • record a likely source/attribution URL for the hash

This would be expensive on the backend, but hey, it's not like we're burning through gazillions of dollars of hardware/compute/FLOPs/electricity just so we can all have a go at "what if Shakespeare wrote the Facebook Terms and Conditions but also as a Charlie XCX song, also please illustrate this in doge memes", and it's not like that has a viable or even sustainable business model right now anyway.

Some reasons why this might be a good idea:

• screenshots of news articles get their source URLs back so you can go read the article (fine, paywalls, whatever)
• it augments community notes or whatever sort of annotation, which apparently are already doing a relatively fine job of checking source material or claims in (a very small absolute number) of posts (albeit potentially high impact ones)
• maybe there's even a bandwidth/storage saving aspect here, who knows

Also fine, if this is an IPFS thing, reply and tell me how this has already been solved.

---

That's it for today. I took a bit of a break, which is totally fine.

How have you been doing?

Best,

Dan

---

How you can support Things That Caught My Attention

Things That Caught My Attention is a free newsletter, and if you like it and find it useful, please consider becoming a paid supporter.

Let my boss pay!

Do you have an expense account or a training/research materials budget? Let your boss pay, at $25/month, or $270/year, $35/month, or $380/year, or $50/month, or $500/year.

Paid supporters get a free copy of Things That Caught My Attention, Volume 1, collecting the best essays from the first 50 episodes, and free subscribers get a 20% discount.