February 2023

121 tweets

Replying to @RobertJBateman and @hemanthkumarak, @siobhansolberg

Yes, ePD is stupidly strict and the exemptions don't have any logic to them apart from "someone at a DPA thought it was OK at some point." For instance, load balancing isn't a user feature, you *could* deliver a shit service without. Yet it's always exempted.
But if you start to include things that the site needs to do in order not just to deliver a service but actually one that's not shit, then you need to exempt more than just that (arguably up to some limited degree of analytics).
At the end of the day, this is pretty bad regulatory structure... It makes harmless things and stupidly dangerous things the same. The requirements should be on the tech to do this right (eg. on browsers).

Thread of 3 tweets

Replying to @JonathanSBlake and @nils_gilman, @NBerggruen, @mangabeiraunger, @NoemaMag

Yes, great article, looking forward to the longer version. Best of luck!
A few words about "The Internet Transition": berjon.com/internet-trans…. I thought I'd never manage to finish writing it because it doesn't fit in my head. Having the cat edit it didn't work. Then I thought no one would read it because it's long and weird. But…
↺ 3
…only two weeks after publishing it, it has now been read twice as much as anything I wrote in the past few years. So many interesting people have reached out, it's been a rare privilege — I just wanted to say: thank you! 💕

Thread of 2 tweets

Replying to @Chronotope

The space is still far too dominated by people who only have a superficial understanding of the digital space. Too much of it is recipes that only work because of specific circumstances, but they have no theory solid enough from which to figure out why and where to go.
↺ 1

Replying to @mberk and @Chronotope

That's absolutely a real thing and I've certainly seen my share. The problem I have in mind in this case is different though. There are things that you can't *imagine*, even humble and truly smart, without a deeper understanding of a system's dynamics.
We're in an era of the digital that I think is coming to a close. The rules of play depend on this specific situation, they aren't "the rules of the Internet." But being prepared for the change means being able to imagine another way to Internet.
↺ 1

Thread of 2 tweets

Replying to @pbannist and @Chronotope

Lol. Deserved sniping aside, the rules of how things work today have been built by and for big tech so both can kinda be true at once! I think the flailing will be in part from a loss of direction as big tech flails.

Replying to @BuffandGrind

Thanks! And I hope I'm right on that bit...
I'm psyched to be speaking at "Privacy, Seriously". It's an online mini-conference convened by the @ICOnews and organised by @Cennydd about, well, taking privacy seriously. Sign up in droves, it's going to rock!
Quoting a tweet by @Cennydd ↗
↺ 1

Replying to @KrisShrishak

"If we have a way to profile you without copying some specific bytes to our servers, that means privacy!"

Replying to @pbannist and @Chronotope

Why would anyone ever resist shitposting, Paul!

Replying to @PrivacyMatters

I think you did :) Previously in the same vein: • berjon.com/privacy-as-pro… • youtube.com/watch?v=xmtjUr… You might find room for disagreement :)

Replying to @null and @MadronaCitizen, @ICOnews, @Cennydd

I don't think we've figured out how to do that online yet. Damn. This is the first use case for the Metaverse that gives me pause.

Replying to @mberk and @Chronotope

Happy to try to explain this better, but it's a bit hard to do in tweet form (and I don't want to write examples down, too).
I'm really sad to see @Coil sunset. They pushed forward when most had given up. Thanks to all of you for the hard work, and I look forward to what will continue to happen in @Interledger!
Quoting a tweet by @justmoon ↗
↺ 3
I would very much like @zeynep to be wrong. Unfortunately, she's usually right.
Quoting a tweet by @zeynep ↗
↺ 2

Replying to @pfoinkle and @radleybalko

Yes, that's on purpose. It's why it's called deceptive design!
At a time when most of the coverage of tech is at best superficial and too often breathlessly gullible, @JuliaAngwin's work has been a beacon of light and renewal. I dearly hope that @themarkup will keep forging ahead, I'm curious to see what she does next!
Quoting a tweet by @JuliaAngwin ↗

Replying to @C4COMPUTATION and @TwitterSupport

Come over to Mastodon, the cookies are better too 😉

Replying to @DanielSolove and @KrisShrishak

This might be relevant to your interests. Sensitive is one dimension, but the argument applies more generally to PETs IMHO.

Thread of 2 tweets

Replying to @dmarti

Replying to @VelenZiga

Literally all I'm doing, like, you checked and all.

Replying to @OR13b

We should also have machine readable rights reservations against this and make sure they're checked and enforced.

Replying to @mcnees and @marthawells1

We are all Murderbot.

Replying to @andwrobs and @OR13b

The thread I cite has one method to trace back to contributions. That's a starting place from which to build audit requirements. I'm not claiming that it's a silver bullet, but if we can police every shop in a country to check that it's paying to play the radio…

Replying to @C4COMPUTATION and @TwitterSupport

That's why I mention Mastodon, at this point it's the only alternative where I'm seeing to sort of "weird overlap" that made Twitter tick. Finding one's way to a good client can take a bit of work, but I'm now seeing far more interactions there on a quarter of the followers.

Thread of 2 tweets

Replying to @privacyfocused

The thing with search engines is that, at the time, there was a fair covenant: they could be relevant from the content they indexed, and in exchange they sent traffic. Everyone grew as a result. Unfortunately, that time has long since gone.
Interestingly, this also shows how detached from the realities of the industry the @iab has become. What adtech needs today is leadership, not some cookie-cutter bravado regurgitating bad takes from ten years ago. Why does anyone pay for that?
Quoting a tweet by @Iwillleavenow ↗
↺ 1
A great frame to parse Google's announcement of #Bard with is @doctorow's recent minting of "enshittification": pluralistic.net/2023/01/21/pot… The core proposition is clear from the UI: everyone gets less say except for Google.

Replying to @faineg

Lol, understatement of the year.
We should standardise SciHub resolution and browsers should ship with it built in.
Quoting a tweet by @faineg ↗

Replying to @PrivacyMatters and @RobertJBateman

There is that, but I think it may also be part of a broader trend to care more about data sovereignty than about privacy. Which of course leads only to a bigger clusterfuck.

Replying to @montezumachavez and @PrivacyMatters, @RobertJBateman

Exactly.
The bottom one tells you everything you need to know about the idea that using people's data can be a "value exchange."
Quoting a tweet by @natashanyt ↗
↺ 10

Replying to @nvbalaji

Sure, you can always reframe a question to trigger specific reactions, especially if you press on money and loss aversion. It doesn't change the fact that no one is willingly entering a value exchange, or that a value exchange can't be if you can't understand what you're giving.

Replying to @sephr

Except that it's not knowable.

Replying to @WebDevLaw

<Bambi Eyes> Pleeeease finish it! Sending over energy!

Replying to @sephr

There isn't a "user journey", that's not something that people have a mental representation of. Most users don't understand that the search engine and the browser are different things, let alone grasp what events constitute a collection trail.
But worse than that, users, no matter how sophisticated, cannot guess at what inferences may be made once the data leaves. There are potentially millions of things that a simple trail can reveal, some of it sensitive, and you simply cannot know what.
So your suggestion amounts to asking people to share something they don't understand, attached to opaque and time-shifted risk, and trade it for a discount. There's no version of that that's a value exchange.

Thread of 3 tweets

A bad implementation of a bad idea executed with so much attention to detail that not even the highly vetted promo is right. It's really tiring to see such dumb gimmicky products get so much funding when we have real problems to solve.
Quoting a tweet by @mmitchell_ai ↗
↺ 1
The glorious future of search is finally here.
↺ 3

Replying to @BuggeErik

This is me being degenerate by my own human self!

Replying to @judyshapiro

Hey, at least I didn't use Microsoft Bob, which was my first idea.
It's been the future of search forever — and forever will be that.

Replying to @ColinHayhurst and @BuggeErik

Hahaha thanks 😁

Replying to @null and @MadronaCitizen

Well thankfully the browser does a really good job of helping people choose a search engine for themselves and not privileging a specific one, in line with putting users first.

Replying to @null and @MadronaCitizen

You wouldn't use a browser to do that.

Replying to @montezumachavez and @alextcone, @Chronotope

Good! ♥

Replying to @montezumachavez and @alextcone, @Chronotope

That will take a while, but we will take them to wide review soon. We'll need your help to ask everyone for input and to give your feedback!

Replying to @nvbalaji

I'm not saying that that's what you're arguing for, only pointing out that people can't exchange what they don't understand.

Replying to @HyperboIeva

See also: writing.

Replying to @__apf__

Don't you have that relationship backwards?
Privacy is an intimately human concern. It is often surfaced to users in several ways, and how data flows out of a product can be critical to its commercial viability. But do we task product people with handling it, as you'd expect?
You know we don't! Instead of approaching privacy as a product concern, we let compliance lawyers run wild with it, often (though there are exceptions ;) with the detrimental effects we all see every day. How can we fix this? I hope to help provide an answer!

Thread of 4 tweets

Replying to @__apf__

"What did you say your name was, assistant person?"

Replying to @random_walker and @sayashk

I think it's in part a problem of empowerment. Unless you know how to make it useful or trust someone who does, only the first part of the title is actually actionable.

Replying to @WebDevLaw

Thanks! Let's hope the pile doesn't collapse!
Oh shit, they trained the AI using men. twitter.com/MovingToTheSun…
↺ 15

Replying to @sethlazar

We definitely know that some are fake, I harbour some suspicions about this one. However asserting that high level of confidence is something that has for sure been seen in the wild, eg. that interaction about the primality of an even number.

Replying to @sethlazar

Emotional states I don't know (doesn't Siri do that?), but commenting on the user does feel like something that you'd expect whatever safety mechanisms they have to try to filter out.

Replying to @sethlazar

It's also unclear which version or variant this is running. They seem to be seeing this as an existential battle and appearing more emotional is probably more engaging. If they haven't removed or diminished that safety yet, they are bound to.

Thread of 2 tweets

Well that's an unfortunate take. We have our work cut out!
Quoting a tweet by @w3c ↗
I'm delighted to see New York Times contributors speaking up together, collectively and effectively, against rampant transphobia in The Times's coverage. I'm proud to see the names of so many friends and former colleagues signing and supporting. nytletter.com
↺ 1

Replying to @gchampeau

Pense aux pauvres 90% qui n'auront même pas 0 euros de plus.

Replying to @sethlazar

That sounds bad! I'm not so worried about manipulation (yet) and more about plain old confirmation bias. These things succeed by giving people what they want and they can by nature hear eg. a political dog whistle or any bias in the question. Feels like a bad combo.

Replying to @sethlazar

Yeah, that's definitely fucking unsafe. It makes me wonder how close we are to stochastic terrorism with this.

Replying to @sethlazar

That's wild. I tend to worry about large scale effects that aren't immediately obvious. "Will actively encourage harming people with emotional blackmail" wasn't part of my threat model.
I subscribe to the paper edition. Who should I give that money too now that I'm cancelling? assignedmedia.org/breaking-news/…
Thank you, @TheOnion, you truly are the beacon of democracy and public debate that others only talk about being. "Institutions with massive platforms like ours must be open to different ways of endangering the trans community." theonion.com/it-is-journali…
↺ 2
This thread is entirely correct, the only change I would suggest would be to extend it to all of algorithmic media, including search and not just social.
Quoting a tweet by @random_walker ↗
↺ 9

Replying to @mattecapu and @Twitter

There's a setting to get it to remind you!
The switch from Twitter to Mastodon is an amazing illustration of how algorithmic curation is a highly impactful editorial process. (I had Twitter set to chronological, but what I see is still heavily influenced by what others RT, and most had the curated timeline.) 🧵
↺ 4
My follows there are overwhelmingly an import of my Twitter follows, and yet I am seeing content I never saw here and finding friends on topics I thought myself isolated. Automating editorial curation does not change its nature.
↺ 2

Thread of 2 tweets

Replying to @ElizabethEnglez

You can get just the same thing over there. It's not different in who you can follow.

Replying to @harriesadam and @_julesh_

I think that varies a lot by country, and perhaps by industry. Requirements that it shouldn't conflict in terms of interests or demands on your time/energy are common from what I've seen, you might need to get permission but I've only rarely seen outright bans.

Replying to @corinneblalock and @OlufemiOTaiwo

Will that be open to people on the Internet?

Replying to @corinneblalock and @OlufemiOTaiwo

Oh! Sorry, I didn't get that. Brilliant, thanks!
Are you working on private retrieval? Could you use a grant? Here's where you apply:
Quoting a tweet by @ProtoResearch ↗
↺ 1
If you're into figuring out questions at the intersection of decentralisation, content addressing, P2P on the one hand and compliance with data and hosting regulations on the other, come work with us in the Decentralized Data Compliance Working Group! github.com/DDC-WG
↺ 1
There's no guarantee that any of these will be accepted, but to give you a sense of the proposals that ✨YOU✨ should be submitting, I just proposed: • A track on dweb policy • A workshop on governance in the IPFS Cinematic Universe • A talk about "What Is The Web?" Join us!
Quoting a tweet by @mishmosh ↗
↺ 2

Replying to @ANderagakura and @ICOnews

Cats can make it… purrrrfect. (Sorry.)

Replying to @divijualsuspect

It happens to be on the "read soon" pile :)

Read the whole thread: 9 tweets →

"If you're not an asshole to the consumer, generally you're okay." — @harrybr giving the kind of near-legal advice that I heartily endorse!
Yeeeees yes yes yes, @epenzeymoog. #PrivacySeriously

Thread of 2 tweets

Replying to @mnot

Congrats!

Replying to @PrivacyMatters and @harrybr

Which is why it should be the central strategy of all compliance people: very little left to do after that 😁

Replying to @InclusiveLucie and @Cennydd

Thank you! I'm really glad you liked it!

Replying to @null and @MadronaCitizen, @harrybr

Eeh, I reckon the Delphic Seven Sages had "Don't Be, Like, An Outright Asshole" as the training level for "Know Thyself". It's easier!

Replying to @sarahtgold and @Cennydd

Thank you! Lit panel too!

Replying to @ru7701 and @CRC4D, @ThomasPHI2

Wow thank you, this is excellent company to be in! Regarding public goods specifically, more than the privacy bits you might like: • The Internet Transition berjon.com/internet-trans… • Stewardship of Ourselves berjon.com/stewardship/ • Capture Resistance berjon.com/capture-resist…

Replying to @othermaciej

LOL — seems unlikely!

Replying to @ethyca and @jason_kint, @random_walker, @JoeGTJones, @ErinRelford

Thank you! 💗💕

Replying to @dietrich and @DeveloperAlly, @glifio

It's a nice intent, but I'm afraid it doesn't scale.

Replying to @montezumachavez and @alextcone, @Chronotope

What's the surprise Luis? Are you visiting?

Replying to @montezumachavez and @alextcone, @Chronotope

Wooooohooo!

Replying to @montezumachavez and @alextcone, @Chronotope

Well when exactly would you be here?