June 2022

253 tweets

I'm not an expert on this: what is the logic in the Star Wars universe that requires bad writing and poorly-filmed chase scenes with children?

Thread of 2 tweets

Replying to @dmarti

But other franchises manage to write for broad audiences without these issues. Eg. Pixar.

Replying to @Chronotope

My god it's so bad. Why? Why?

Replying to @C4COMPUTATION and @Collective_CI

On a rolling basis: I assume that this means online-only?

Replying to @pbannist and @alextcone, @robleathern, @caitlintackles, @swodinsky

CRIOSPHINX: Context-Respecting Innovation with the Objective of Safeguarding the Privacy of Humans In Networked eXperiences. Wikipedia: "The task of [criosphinxes] was to hold back the forces of evil." en.wikipedia.org/wiki/Sphinx

Replying to @pbannist and @alextcone, @robleathern, @caitlintackles, @swodinsky

Replying to @kkomaitis and @Chronotope

Good choice!

Replying to @MaxGendler and @swodinsky, @pbannist, @alextcone, @robleathern, @caitlintackles

Replying to @Chronotope and @dmarti

I actually don't mind too much that there's a kid in it (though very few writers can do kids well) but those chase scenes with the mercenaries and then again Obi-Wan trying to catch her in the market? It looked like they kept missing her on purpose.
Retrofuturism: some notes on @LMSacasas's recent piece about how our tech world makes us feel stuck in the past, as if nothing can change. berjon.com/retrofuturism/
↺ 11

Replying to @scodtt and @Chronotope, @dmarti

Totally! Bring me more Loki!!!

Replying to @Chronotope and @scodtt, @dmarti

Exactly! There's a whole galaxy of wild worlds but the only thing that matters is one family. I guess in a way that belongs to our time.

Replying to @BrettFrischmann

Thanks Brett!

Replying to @null and @CJAMcMahon, @LMSacasas, @PetterBB, @tainab

I'll look those up but if you have specific paper recommendations I'd love to hear!

Replying to @IEthics

Thank you for all the tweets!

Replying to @beccanalia

Thank you!

Replying to @LMSacasas

I have been getting only half my mentions all day for some reason. I hope you don't hate it!

Replying to @LMSacasas

Thank you! Glad you liked it.

Replying to @MissIG_Geek

It is, and we should. But the problem is that libertarian paternalism isn't just the ideology of nudging, it's the moral framework of the people who see themselves as doing good in tech (esp. at Google). Until we replace that with better roots, we'll be treading water.

Replying to @null and @blakereid, @nabihasyed, @LMSacasas

Thank you!

Replying to @LMSacasas

I don't know about better articulated, it's hard to beat Saturn! But overall I see the theses as related and in a few cases I could have picked a different order with limited changes.
The thing that I found particularly potent in your piece and that made me want to write this is how this feeling of stuck time brings together a number of disparate strands. I have felt that things are stuck for a while but I saw that as a technologist: stuck progress.

Thread of 3 tweets

Replying to @farzanehbad

Hahaha very fair point!

Replying to @null and @CJAMcMahon, @nabihasyed, @LMSacasas, @PetterBB, @tainab

Thank you!
“Interestingly, it turns out that some of the ad experiences that consumers find the most annoying are also bad for the environment.” And it so happens that this is a good time to fix this because we're getting rid of identifiers. 🧵 digiday.com/media/the-ad-e…
↺ 2
You can get better privacy, better publisher revenue, and a more eco-friendly ecosystem by batching and switching to matching markets, which offer a well-defined way of running auctions.

Thread of 4 tweets

“For many years, the accepted common knowledge was that the only way to make money online was to adopt Google’s model—that surveillance capitalism was the way to go if you wanted a profitable and sustainable and scalable business. We have proven that there is another path.”
Quoting a tweet by @mario_gug ↗
↺ 3
I'm very curious to see if Google will make the case that working in the WICG meets its @CMAgovUK commitments to proceed through the appropriate W3C process.
Quoting a tweet by @ANderagakura ↗
↺ 2

Replying to @null and @blakereid

Totally agreed, I think the fact that there isn't a market for privacy is established (and there shouldn't be). Most of the VC-supported "boom" in privacy startups is for companies that are either security, compliance, or notice-and-choice stuff.
However, what I found interesting here is that this is a long-established company that is now *growing* and has a business model. I think there's a trend, and not so much because of privacy but just because more people understand you have to pay if you want quality.
People use privacy as a comparator with Google because, well, it's easy for anyone to be better at privacy than Google. But the reason those products are interesting is usually because they're just flat out better. That's particularly obvious in search & email.

Thread of 3 tweets

Replying to @null and @blakereid

I beg to differ on that point. DDG used to be worse, but that hasn't been the case in a while IME. Partly because its index has improved, partly because Google search is really sinking, notably in UI (but not only).

Replying to @null and @blakereid

Yeah, that used to be my experience, but recently it's gone the other way where I'm on someone else's device and have to switch away from G because it's subpar and has a taxing UI. And for image search, DDG is a lot better. Have you tried Neeva? I've had a good experience.
TFW when a major draft federal privacy bill is called ADPAPA. #ADPAPA
Quoting a tweet by @swodinsky ↗
↺ 1

Replying to @Chronotope and @swodinsky

Yeah, it looks like they are trying to make well-intentioned carve-outs but failing. This looks fixable though. Their definition of third-party collection that doesn't cover Google is more worrying and reads more like lobbyist intervention though.
This is a bad definition of "third-party." Let's hope we can get it fixed.
Quoting a tweet by @swodinsky ↗

Replying to @dmarti and @Chronotope, @swodinsky

Limits to preemption that keep the CPRA alive and given states general options for other stuff that will be dropped.
↺ 1

Replying to @thedooosh and @JeromySonne, @swodinsky, @Chronotope

It's been the thorn in a lot of "just use contextual" proposals. Keeping the ability to measure is crucial in all cases.

Replying to @swodinsky and @ewlamy

I mean everything is an algorithm, possibly. There's a reason why people usually focus on automated decision-making, it's easier to specify. An ad auction (typically) wouldn't be covered (with caveats for the kind of dirty tricks some folks play, eg. G), but...
...bidding decisions likely would. I'm generally skeptical of transparency and impact assessment measures, but a requirement to explain what data (including "anonymised") goes into the decision explained in terms of *where it comes from* rather than vague categories would rock.
↺ 1

Thread of 2 tweets

Replying to @dmarti and @fatemehx2, @Chronotope, @swodinsky

Yeah and the point of the GDPR was to protect people's privacy and yet here we are.
↺ 1

Replying to @fatemehx2 and @dmarti, @Chronotope, @swodinsky

Fucked around, finding out!

Replying to @ewlamy and @swodinsky

Or even just the algo that each DSP uses in the auction.

Replying to @ewlamy and @swodinsky

Right, it takes some solid testing to know what is significant in a given model. Though of course if it has impact on people that's probably not a reason to not do it!

Replying to @montezumachavez and @jduballreports

Too much for a Friday!
When you see industry orgs pushing absurd positions, this is what's happening: they're giving cover to their worst members. If you're not one of the worst members and you don't speak up, it'll haunt you when you need your org to be credible.
Quoting a tweet by @dmarti ↗
↺ 2

Replying to @Cyn_K and @swodinsky, @thoward37

To be fair we should have regulated spreadsheets, but for other reasons 😁

Replying to @braedon and @swodinsky, @Chronotope, @daphnehk

I guess that depends on the procreation method?

Replying to @braedon and @swodinsky, @Chronotope, @daphnehk

Just doin' my job, sir.

Replying to @C4COMPUTATION and @sfiscience, @Harvard

Do you know if it will remain available after it has streamed? This is extremely my area of interest.

Replying to @C4COMPUTATION and @sfiscience, @Harvard, @YouTube

Brilliant, thank you! That way I won't regret being on another continent at that time.

Replying to @swodinsky and @Chronotope, @AlanInDC, @daphnehk

Lawyers have parties to a contract, and anyone else is a third party. The normal tech understanding is similar, not a contract but the parties to an interaction are the person and the first party, others are third parties.
Adtech people like to confuse things because otherwise they'd have to accept that they logically have fewer rights. Eg. the Criteo/SWAN antiprivacy activists claim that Disqus is a first party and a publisher. Laughable but very legacy adtech.

Thread of 3 tweets

I think we're going to have to have some words about the Spelling Bee's metaphysical commitments.

Replying to @EvanSelinger

Do not tempt me with a good perdurance!
Any sufficiently advanced technology feels like an extension of yourself; if it feels like magic, look for who's running the show.
↺ 9
Seriously, though, what do you need Shoshana? Are you generally looking in the direction of unfucking the computery things? That rabbit hole has rabbit holes; some ideas in there:

Thread of 2 tweets

And then these. There are also articles... And sign up to the Convivial Society. Let me know if you want to chat, always happy to talk with fellow travellers!

Thread of 3 tweets

Replying to @null and @benharnett, @swodinsky, @MaxGendler

Credit where due: it's a recommendation from @EvanSelinger (as are a few other titles in there).

Replying to @swodinsky and @MaxGendler

There's just too much 😭 I wish I could get paid to read/think/write instead of mostly to be in meetings…

Replying to @HalSinger

There's real risk Hal, if you skip school you might *checks notes* end up being hired as a VP at The New York Times.

Replying to @kkomaitis

You know there are other search engines, right?

Replying to @profmadison and @LMSacasas

Thanks! I think domestication can bring together notions of nudging, deskilling, and convenience on one side, and the issues from the loss of robustness and diversity in engineered ecosystems on the other. I need more work on this to make it crisp.
Slightly further afield, domestication also comes with a host of ills and illnesses for humans, animals, and plants — I wonder if the digital equivalent has comparable mechanisms.

Thread of 2 tweets

Replying to @profmadison and @LMSacasas

Absolutely, and the ties to power and statecraft too. Scott focuses a fair bit on the evidence that people were forced into agriculture (because it's not generally better for them) and how grain and taxes work together.

Replying to @mariafarrell and @jilliancyork

I think I saw one the other day with the loose change at the back of the couch, but it was kind of far and with those gross crumbs.

Replying to @mariafarrell and @roessler

Wait, if that's a training you give I will look for budget!

Replying to @kkomaitis and @jilliancyork, @BikesNBukes, @mariafarrell

Where is the preorder button?
I'm sorry to have to share that Mochi, one of our beloved cats, passed away last night. He was the sweetest of cats and barely over one year old. Our time together was too short. A short thread of memories to share his fluffy kindness.
↺ 1
Mochi was very dedicated to building a digital ecosystem that supports sustainable media and a healthy society. He spent a lot of time in meetings working towards those goals.
He was particularly well-read, even by cat standards. He had a particular fondness of the James C. Scott classic "Sleeping Like A State."

Read the whole thread: 8 tweets →

Replying to @ewlamy

Thanks Eric.

Replying to @swodinsky

Thank you. Yes, he really was.

Replying to @mcnees

Thanks Robert. ❤️

Replying to @null and @avadacatavra

Thanks, indeed.

Replying to @_aurelienfabre

Merci!

Replying to @TomKemp00

This is a quality meme, Tom.
↺ 1
This shouldn't hide the fact that the warts do need to be addressed, but the draft both significantly improves people's privacy and helps publishers "keep data where the media is" which is a huge win. It is set up to rectify market power imbalance.
It's hard to tell if it has any chance of moving forward or passing, but if it doesn't we should learn from this draft as it provides a credible architectural blueprint for better privacy law.

Thread of 3 tweets

Replying to @SamuelStolton

Thank you!

Replying to @ehsanakhgari

Yes, it is, thank you.

Replying to @alextcone

Thanks Alex.

Replying to @mariafarrell

Thank you Maria ❤️. Yes, I'm not sharing the pics where he's with the girls but he was absolutely adorable a childhood companion.

Replying to @BuggeErik

Thanks Erik, he was very dog-like in fact! He followed us in the street, if we ran in the backyard he'd run around too, he tried to bring back balls.

Replying to @montezumachavez and @alextcone

Thank you Luis, hugs.

Replying to @_abialas

Thanks Anna; I included the yoga picture because I remembered how much his face had cracked you up.

Replying to @null and @podopie

A fortress of books and warm feet — it's important.

Replying to @null and @podopie

Thanks, yes, they do. We need the fluff.

Replying to @privacypen

Thank you Angelique.

Replying to @LourdesTurrecha

Thanks Lourdes, that's kind.
Intéressant: dans la 1ère circo des Français de l'Étranger le candidat centriste-écolo, Gérard Michon, n'appelle *pas* à voter pour @RolandLescure (qui a passé la pandémie planqué à Paris). Il serait logique que le report se fasse sur @florencergr de NUPES, qui peut donc gagner.

Replying to @null and @TrialByTruth

Thank you!

Replying to @fabricedesre and @RolandLescure, @florencergr

À ton service!

Replying to @chris_bail

Lol, the *real* boss level is how many "email_bankruptcy_20xx" mailboxes you have and what the unread count is on those 😁
There should be a Bullshit Beat and Aram should be on it.
Quoting a tweet by @Chronotope ↗

Replying to @willseth

Thank you!

Replying to @null and @amyvdh

Thanks Amy ❤️

Replying to @Iwillleavenow

Thank you, that's very kind ❤️

Replying to @swodinsky

I was thinking you get to be the EIC of the whole thing.

Replying to @daniellecitron

Thank you Danielle 💗

Replying to @rdeltour

Thanks. Yes, I hope that too! He used to sit right at the beginning of the forest and stare into the trees. He's meant for that.

Replying to @LAM_Barrett

Thank you Lindsey.

Replying to @ariezrawaldman

Thanks for these kind words. Yes, indeed.

Replying to @alliecmorgan

Thanks Allie!

Replying to @null and @blakereid

Thank you.

Replying to @null and @prushforth

Thank you 🫂

Replying to @innovimax and @tom_yamahito, @_dmaus, @gimsieke, @stevenpemberton, @markupdec, @xmlprague

A lot depends on how big you need it to be. It doesn't have to be as complicated as berjon.com/w3c-governance/ for instance! Funding is the hard one — @tobie might have some good insights there.
↺ 1

Replying to @reidmore_online and @wendy_a_reid

Thank you Wendy, indeed on both counts!

Replying to @stevenpemberton and @tobie, @innovimax, @tom_yamahito, @_dmaus, @gimsieke, @markupdec, @xmlprague

I'm not sure that labelling "this is a standards body" and "this isn't" is a good way to solve this. What are the requirements, what does a CG provide and not, etc. A CG is probably enough for open source+specs+patent commitments for instance.

Replying to @null and @fran_mady

Thank you Fran!

Replying to @null and @benharnett

Thanks Ben 😔

Thread of 2 tweets

Replying to @Aurorabfar

Same!

Replying to @null and @MsBrandy99

True, I'm not absolving Amazon of *waves around* but people know which shopping experiences will wreck their inbox and which won't!

Replying to @LMSacasas

I don't know if it's the default setting, but IIRC Aristotle mentions it as something tyrants use in the Politics, so it's been around. I liked @henryfarrell's idea of "common sense attacks" on democracy. It opens the door to thinking about protective measures.
Unfortunately, it's also well-known that children raised to think that paying $47 is normal for an article have the worst life outcomes. Teach your kids about the heroism of SciHub.
↺ 1

Thread of 2 tweets

Replying to @coolharsh55

Thanks, I know 😁 But $47 is a new low I thought!

Replying to @peligrietzer

Shit, you have a point. We can't just burn it all down.

Replying to @kkomaitis

Replying to @alextcone and @montezumachavez, @kkomaitis, @MaxGendler

For Luis!

Replying to @mariafarrell and @kkomaitis

Replying to @mariafarrell and @kkomaitis

Next time we do London. Or all to Lesvos!
Cool jobs if you want to improve privacy for Californians!
Quoting a tweet by @CalPrivacy ↗
↺ 4

Replying to @swodinsky

↺ 1

Replying to @null and @MadronaCitizen

On the contrary, that's what's exciting about it!

Replying to @MaxGendler and @swodinsky

I'm in Brussels too and I can say that this week!

Replying to @MaxGendler and @swodinsky

Max, don't forget I'm Australian. youtube.com/watch?v=v2AC41…

Replying to @btsavage and @swodinsky, @eric_seufert

Agreed. Privacy isn't about what people perceive, people perceive all kinds of crazy things! It's still far from perfect but we do try to document this in w3.org/TR/privacy-pri….

Replying to @eric_seufert and @btsavage, @swodinsky

Why are you bringing NBs into this? It's well established that individuals aren't in a position to exercise their agency so I'm not sure what we get by pretending that people are in a position to "decide." As cited above we're aligning on pretty clear definitions.

Replying to @eric_seufert and @btsavage, @swodinsky

Yeah, same.

Replying to @chris_bail

↺ 1
If you're at #EDPSconf2022, come say hi!
"To get effective enforcement, it has to pay off to comply. Right now it pays off not to comply." @maxschrems, #EDPSconf2022 Max further explains that we have to learn from sociology of law why people comply in other domains so that we can make it work for privacy.
↺ 10
@PaulNemitz wants to involve the tech community more into privacy. I think that is a promising avenue, but the bridge between legal and architectural enforcement does not yet exist. Also, the tech matters, but so does economics. #EDPSconf2022
@PachlUpa talks about how people give up in the face of pervasive surveillance. This learned helplessness is not the society that we want to live in. #EDPSconf2022
@maxschrems "In the privacy sector, most people don't believe in privacy." #EDPSconf2022 There is privacy as compliance, trying to do as little as possible, and privacy as product/strategy, aiming to deliver better for your users and improve your business. Can these be aligned?
↺ 4

Replying to @podehaye

Hi! I presume you don't look like your avatar?

Replying to @podehaye

I'll look for you! I have white pants and floral shirt.
"The case is pending in Ireland." #EDPSconf2022

Replying to @tjmcintyre

Of which year? 😁
@maxschrems rightly points out that the bureaucratic parts of the GDPR come from the compliance community, not the privacy community. #EDPSconf2022
↺ 4
@vestager specifically asks who will retain access to data after the phase out of third-party cookies? Analysing this correctly will require looking at timing attacks, probabilistic matching, tracking by user agents, in-app browsers... It won't be straightforward. #EDPSconf2022
↺ 1

Replying to @jesslynnrose

A whole paragraph at once, huh? Lucky you!

Replying to @null and @sivavaid

But the pages run on time, Siva. (Is the talk recorded? This should be said more!)
Jane Horvath says she has hope that there is a window to get a US federal privacy law. #EDPSconf2022 #ADPPA

Replying to @justinhendrix and @sivavaid

Face à l'arrivée imminente de l'Étoile Noire, l'État met en place un numéro vert.
Quoting a tweet by @LeHuffPost ↗

Replying to @sara_ann_marie

The problem is that giving feedback on the feedback will make you "opinionated." I've been wondering if we should have feedback panels like we have interview panels, structured to try to weed out this kind of bias.
@VeraJourova speaks of coming new regulation complementing the GDPR: "Perhaps even ePrivacy!" #EDPSconf2022

Replying to @montezumachavez and @alextcone, @maxschrems

Luis, this is a serious question: how good are Colombian fries? If they are not good, what can replace that? I ask because I would love to repeat this, and I think Alex too, but we had really good fries 😋

Replying to @BuggeErik and @maxschrems

Yes and no. The two are related, but product or strategic privacy seek to make your business better whether you have to or not. We have a hellscape in part because we've left compliance lawyers run privacy, which is like having accounting drive your revenue model.

Replying to @montezumachavez and @alextcone, @maxschrems

Wait, but it doesn't have to be fries! Tell me about arepas!

Replying to @montezumachavez and @alextcone, @maxschrems

That is a good idea.

Replying to @montezumachavez and @alextcone, @maxschrems

OK I'm hungry already.

Replying to @kkomaitis

Hmmm, I disagree on pretty much all of that!

Replying to @kkomaitis

Both. It's an active domain in which there is much to do and it's a stable institutional form. What we exhausted twenty years ago is "give stuff away and somehow magic, yay!"

Replying to @kkomaitis

That, absolutely agreed. But I don't think that that's what they're doing. I know at least one smart person in there.
Enjoying a little impromptu celebration of Iceland's national holiday! 🇮🇸 🍰
↺ 1

Replying to @doriantaylor

I think that's what the icing is.

Replying to @doriantaylor

Not so much for the shark 🦈.

Replying to @BuggeErik and @maxschrems

Aah, sorry, my bad, I had misread you. Yes, totally agreed! We're beating people up with carrots, it doesn't work 🥕!

Replying to @alextcone

Absolutely, I've seen great work on this (thinking of you @fran_mady!). But I have been wondering if privacy shouldn't formally be made a product function, with the compliance team one of the stakeholders. It might be worth a Blog Post™.

Replying to @alextcone and @fran_mady

It should definitely be a CEO issue in terms of strategy and discussion, but I still think that operationally it's a product thing. And yeah, we should all be more Dorthe!

Replying to @alextcone and @fran_mady

💯 It's hard to build and hard to keep but so fucking worth it. It's a blessing to have been in such a team and not something I could forget.

Replying to @MaxGendler and @montezumachavez, @alextcone, @maxschrems

I think that with @BuggeErik we've established that it's about carrying a big carrot 🥕!
This, and I also sometimes think of the kind of machine cognition we might have if AI/ML weren't mired in epistemic individualism. The idea of intelligence without an epistemic community is absurd.
Quoting a tweet by @billt ↗
↺ 3

Replying to @hypervisible

Ooooh, yes please!

Replying to @Leo_Bich

This looks amazing. Will proceedings be published?

Replying to @AnupamChander

I'm seeing this late, but definitely got your back! I hope you don't let it get to you at all.

Replying to @mtomweb and @RickByers, @mnot

Infrastructure provision is always a problem, infrastructure provision that rides on investment by big companies is always going to be worse, no? We have a bad business model for browsers (mostly selling users to search) and none for engines. That might be the first fix?

Replying to @mtomweb and @RickByers, @mnot

Indeed, but tricky doesn't mean unsolvable :) 450m is a lot of money, but at the scale of the Web it's not *that* much, and we only need so many engines. I'd say a global budget of 2-3bn for all engines should be enough?
Right now they come from the original sin of the Web: pretending that money doesn't exist so that the only option is to monetise people in variously icky ways. I don't think there's a purely technical solution, but we can look for technical options with minimal regulation.
It's a much longer write-up than a tweet, but an interoperability mandate on search and small revshare to browsers and engines could get the money from essentially the same sources (so we're not breaking the bank) but without the downsides.

Thread of 3 tweets

Replying to @mtomweb and @RickByers, @mnot

Ha, thanks — it's on my pile of things to write. It's a big pile, sadly, so no promises on timing. I have most of the argument lined up, it's a matter of making it digestible.

Replying to @RickByers and @mtomweb, @mnot

By search monetisation, do you mean monetisation of search or using search to monetise browsers? The former I agree isn't icky, the latter I think that if browsers picked based on what's best for people, they'd pick differently — the difference is the ickiness in my mind.

Replying to @RickByers and @mtomweb, @mnot

Hard to sketch in a tweet, but at a very abstract level we can look at "taxing" search such that a small fraction goes to browser and engine, per hit. Works best with a search protocol. Need registration, contract terms for all parties.
This also needs anti-fraud measures, the ability to stream tiny payments, and probably some form of privacy-preserving advertising. The latter is coming, payments has ILP. We need some regulation but not much.
I'm skipping over a ton of details but I think there's a possibility. My hope in writing it up is that people can help simplify it to the point that it works. We'll see!

Thread of 3 tweets

Replying to @RickByers and @mtomweb, @mnot

True, but that doesn't prevent search from becoming a closed platform. I see a (short-term) future in which using a browser is something the weird Linux people do; everyone else just has webviews in a search app. Is that still the Web?
↺ 1
We thought we would stay in the US at least a couple more years as my spouse is doing an MA here, but that's looking like a worse decision with every day that goes by.

Replying to @null and @benharnett

It really fucking does.

Replying to @talitacpessoa and @tatwittando, @TrialByTruth

100% — the problem is that it becomes hard to do policy when you're dealing with failed institutions. I'm not saying that that's an excuse, but it's the barrier.
This is not a drill. People will die when you get this wrong.
Quoting a tweet by @evacide ↗
↺ 34
There is currently a draft federal privacy bill that has advanced farther than any before it. If strong enough, it could outlaw selling location data used in hunting down people who get abortions. You might want to push your Congresspeople, notably @SenatorCantwell.
Quoting a tweet by @motherboard ↗
↺ 3
Individuals cannot protect their privacy. Notice & consent cannot protect privacy. We need strong, effective regulation.
Quoting a tweet by @zeynep ↗
↺ 3
Bonne fête aux amis québécois!

Replying to @epc_angela

This is an amazing piece, thank you. As for "what's next," the scary part is that it seems altogether too predictable.

Replying to @ariezrawaldman

Is it not just a way for companies to pretend they want to do something, while knowing very well they won't?

Replying to @mdennedy and @montezumachavez, @dansereduick, @peer_rich

Also, Canadian swear words should go a long way in France.
It me, too.
Quoting a tweet by @EtheHerring ↗
An inspiring profile of @OlufemiOTaiwo, well worth your time. grist.org/culture/olufem…
This might be the best summary of one of my core concerns.
Quoting a tweet by @ArthurB ↗
↺ 4

Replying to @braedon

Ha, I had the same idea a few years back, tried to get us to do it for the Privacy Project. I think that with just a hint of logic you could get some really fun ID graphs.

Replying to @swodinsky and @JeromySonne, @BradAT, @Therran, @Chronotope

The little-known Spiderman Corollary is that you can't have a lot of responsibility if you don't have much power. Once someone else's script loads, publishers have almost no power, and they have no visibility into what makes money to pick what's worth doing.
↺ 1
My point here isn't to absolve pubs from responsibility: on the contrary, pubs should be given the power to fix this. We'd get innovation at the edges, a much cleaner system, and a clear way to blame people when things go wrong.
↺ 1

Thread of 2 tweets

Replying to @JeromySonne and @swodinsky, @BradAT, @Therran, @Chronotope

No worries, I didn't get to see whatever ghastly insult it is you made :-D

Replying to @JeromySonne and @swodinsky, @BradAT, @Therran, @Chronotope

We're charging, we're charging. But we need a fair bit more than transparency, we need to be running targeting decisions pub-side. That's the only way we can be accountable for what's going on and build things that put our readers first.

Replying to @JeromySonne and @swodinsky, @BradAT, @Therran, @Chronotope

I think that this is a case of It's More Complicated™. The NYT can audit _some_, but frankly once you get to a loaded creative there's really not much we can do. And if we can't, most publishers can't. If all the bad things have to be decided by the pub, then blame is clear!

Replying to @swodinsky and @JeromySonne, @BradAT, @Therran, @Chronotope

Yeah, but if pubs don't have enough power to pick gnarly shit or no gnarly shit (we have some, just not enough) then everyone gets some degree of gnarly shit. And if you complain, in so many cases my best answer is "it comes from open market programmatic ¯\_(ツ)_/¯!"

Replying to @JeromySonne and @Chronotope, @swodinsky, @BradAT, @Therran

I agree with this. Markets can be fine when they are set up right and governed properly. Insert the usual about markets ≠ capitalism :)

Replying to @JeromySonne and @dmarti, @Chronotope, @swodinsky, @BradAT, @Therran

Contextual + SDA + coarse geo would be great for pubs, too. It's not us who don't want that world :)

Replying to @pbannist and @robleathern

Also I'm just as clueless as Paul about sportsball so I have no idea how to address this.

Replying to @swodinsky and @Chronotope, @dmarti, @JeromySonne, @BradAT, @Therran

We have a pretty good idea. I'm pretty sure that most intermediaries don't.

Replying to @swodinsky and @Chronotope, @dmarti, @JeromySonne, @BradAT, @Therran

Most of the middlemen don't bring any actual value, which makes it hard to explain… If we could trim down to just the useful middlemen, this could all fit on a diagram and make sense.
↺ 1

Replying to @swodinsky and @JeromySonne, @Chronotope, @dmarti, @BradAT, @Therran

Agreed — pub-owned, with user accountability. The problem with self-serve is scale. But we could federate and do automated buying much better than the current system.

Replying to @swodinsky and @Chronotope, @dmarti, @JeromySonne, @BradAT, @Therran

I meant a useful one :)

Replying to @swodinsky

You don't get it, it's because you're SURFING.
This is doubly good for being about adtech. Now I really want to read Standage 1998, which is about The Victorian Internet (trnmag.com/Bookshelf/The%…)!
Quoting a tweet by @swodinsky ↗

Replying to @swodinsky

Hmpf, I dunno. Where's the Zeroth-Party Data in there, Shosh? Where?

Replying to @swodinsky

Surfing!

Replying to @swodinsky and @DinaSrinivasan

Rand has any number of issues, but they're pretty damn good at the research.

Replying to @kelsiemvn and @wef

I don't think that you can define that outside of a specific context? Given the values of an institutional arrangement, good governance will robustly, sustainably further those values.

Replying to @chris_bail

That tee is pretty lit, too.
I don't know, I'm starting to get this vibe that having a bunch of unelected, unaccountable folks running a whole country isn't great.
Quoting a tweet by @stevenmazie ↗
↺ 10

Replying to @random_walker

Privacy-friendly advertising is a thing, I hope you include it! It's a growing space, and there are companies like Kubler doing great work.

Replying to @random_walker

That's because I can't type a six-letter name without a typo — it's Kobler, kobler.no. If your Norwegian isn't up to par, you should chat with @BuggeErik.

Replying to @ntnsndr

Or begging for the end of the Supreme Court.
Some very quick thoughts. From a privacy perspective, using some limited data from sites {A,B,C} to advertise on site D could be fine *if* D doesn't actually learn anything and the person can never be recognised across contexts at all. Some form of opaque reusable SDA.
From a strategic perspective, the fact that data from one context can be used in another isn't just a privacy problem, it's really bad for publishers too when they can't control what other contexts and purposes it gets reused in.

Read the whole thread: 7 tweets →

Replying to @megyoung0 and @datasociety

This is brilliant!

Replying to @humanpropensity and @larakiara, @pbannist

Absolutely agreed. We could use ML but it would be better to be able to develop and share models inside leagues.

Replying to @humanpropensity and @larakiara, @pbannist

I would still want to pick who to share with. Speaking of course completely personally, I would be comfortable with eg. local news, niche quality pubs, or useful services that can't have strong signals (eg. a timezone site), but not antivax disinfo or whatnot.

Replying to @pbannist and @humanpropensity, @larakiara

Right — I don't think that this is about big vs small pubs. It's a question of signal strength and how well you know your audience.

Replying to @humanpropensity and @larakiara, @pbannist

The complexity might be manageable. Something "like" first-party sets plus w3.org/community/webm… could already go a long way.