Replying to @null and @nitot
Merci!
January 2021
352 tweets
Replying to @null and @ManohlaDargis
The real magic is people who you grind to a pulp for fuel.
Replying to @mraginsky and @mikeandallie, @C4COMPUTATION
Hahaha, this is the truth indeed. That book defeated me twice, I've been wondering if I should try a different approach, maybe a slow chapter per month, a reading group, or something.
Replying to @_developit
It's weird, right? Not sure what happened. All I seem to get is a Kardassian telling me about migraine pills.
Replying to @_alastair
Technically, I'm not sure that you need to be sober for the vaccination itself.
♥ 1
Replying to @_developit
Ingenious!
Congratulations to @k_johnsn on being elected to the AMP AC, representing NYT! Kelsey has seen first-hand just how painful AMP can be to work with, I look forward to seeing her help lead the Web out of this impasse.
blog.amp.dev/2020/12/14/amp…
♥ 8↺ 1
It's Needle time again!
nytimes.com/interactive/20…
♥ 4
Replying to @mikarv and @tim_libert
And from the time before!
I'm losing track of the proposals Google has to turn the web into a strip mall they own and operate: AMP, CWV hiding AMP, Web Bundles, Google News... Now pitching Showcase as a strip mall model to address issues from previous strip mall models.
Maybe try something new instead?
Quoting a tweet by @emilybell ↗
♥ 38↺ 7
I think this is how you imagine revolution works if your reality is social media. t.co/BOUoch6Xow
♥ 10↺ 1
I understand why one would not be subject to the full responsibilities of publishers for just hosting someone else's content — but can someone tell me why you wouldn't be considered to have editorial control the second you add a recommendation system on top of hosting?
♥ 28↺ 5
It's now official: Dry January to start next month.
♥ 34↺ 2
Replying to @alex_kuhner
Ah! I should try to make it back there next time!
Replying to @null and @MadronaCitizen
Actually in this case I like that they gave room to the pics. They show something that too many didn't think could happen.
Replying to @null and @MadronaCitizen
I always wonder to what degree brands that advertise on YouTube take into account the risks that come with it being one of the most brand unsafe environments on the Internet.
Quoting a tweet by @calm ↗
♥ 7
Replying to @swodinsky and @Chronotope
None of this would exist were it not for YouTube and Facebook, both of which have everything driven from adtech. And none of the nazi press would exist without programmatic and programmatic fraud.
♥ 31↺ 4
Replying to @ALadyInKuwait
But this is not just about today, YouTube has been a dumpster fire of nazism and conspiracy theories for years now. Advertising there means taking exactly this kind of risk!
Today's coup attempt, and much else in the current climate, can be traced directly to deep failures in the governance of tech.
We shouldn't expect those companies to fix this. They've shown themselves to fall well short of the challenge.
It's hard. The turbulence isn't over.
Quoting a tweet by @Chronotope ↗
♥ 42↺ 9
Replying to @Chronotope and @taubeneck, @acfou, @swodinsky
I would add: building a system at this scale, particularly one with characteristics long known to be problematic (eg. privacy), without pausing to think about what predictible consequences it may have isn't "unintended consequences", it's flat out negligent irresponsibility.
♥ 5
Replying to @Chronotope and @taubeneck, @acfou, @swodinsky, @monteiro
Yeah, that book is awesome. @monteiro was kind enough to send me a whole bunch of copies, lots of people at The Times have read it thanks to that!
♥ 3
Replying to @swodinsky and @caitlintackles, @taubeneck, @Chronotope, @acfou
Indeed, there are good people, but the governance failures largely dominate. I mean, AdChoices? Even right now, IAB Europe is working on a code of conduct that benefits itself potentially over any of its members.
♥ 2
Replying to @tobie and @rowan_m
I haven't transferred to the WICG pending discussion of who owns it.
Congratulations to @rhiaro, @hober, @LeaVerou, and @sangwhanmoon for their (re)election to the @w3ctag, as well as to @hadleybeeman for her re-appointment!
This was a tough election, there were more candidates that I wanted to pick than available slots. I hope they engage!
♥ 17↺ 4
Replying to @megancgraham and @Greg_Asquith, @acfou, @mchowla, @pbannist, @swodinsky, @Chronotope, @taubeneck, @DisinfoIndex
These intermediaries are making money from copyright violations that were reported to them. How is this not a lawsuit?
♥ 4
Replying to @acfou and @megancgraham, @Greg_Asquith, @mchowla, @pbannist, @swodinsky, @Chronotope, @taubeneck, @DisinfoIndex
Yeah, I don't care about the sites, I care about the companies that make money serving them after being informed of the issue.
♥ 4
Replying to @mchowla and @megancgraham, @Greg_Asquith, @acfou, @pbannist, @swodinsky, @Chronotope, @taubeneck, @DisinfoIndex
Aiding and abetting. They know, and they profit from it.
♥ 1
Replying to @mchowla and @megancgraham, @Greg_Asquith, @acfou, @pbannist, @swodinsky, @Chronotope, @taubeneck, @DisinfoIndex
They act. They are told that they support a copyright violator. They switch the faucet off. Then they decide to switch it back on.
♥ 1
The UK’s competition authority, @CMAgovUK, has opened an investigation into Google’s “Privacy Sandbox”. This is a fascinating topic at the intersection of privacy and market power. If they act, it could be either great or terrible for consumers. 🧵 gov.uk/government/new…
♥ 49↺ 29
What is the “Privacy Sandbox”? It’s the name of a loose bundle of proposals from Google to improve privacy in advertising. Before diving in, a few misconceptions about the PS need to be addressed.
♥ 3↺ 1
❌ The Privacy Sandbox is a "fait accompli".
No, it is being actively discussed and nothing unilateral has yet happened. Chrome has done unilateral in the past, but not this time (at least not yet). The discussions in standards groups are very active, with more to come.
♥ 6↺ 1
Replying to @mario_gug and @CMAgovUK, @johnnyryan
Hey, if it means that the ICO enforces purpose limitations I'm happy!
♥ 1↺ 1
Replying to @null and @MadronaCitizen
That, except it's like Renault and Siemens doing the shaving.
Replying to @null and @pes10k
I don't think that they have such good opsec that the FBI can't find them some other way!
Replying to @Iwillleavenow
I have to say that seeing you livetweeting CATS in the middle of [gestures broadly at all the things] is quite something 😁
♥ 1
Replying to @slightlylate and @_alastair
Also stores have built-in monetisation. That works differently on the web but they could end up being demonetised just the same while still installable.
♥ 1
Indeed we need antitrust so that the moderation decisions of a few platforms carry far less weight, but also well upstream of that so that their editorial recommendations don't create such potent harms from speech in the first place.
Quoting a tweet by @doctorow ↗
♥ 21↺ 4
Replying to @BrendanEich and @brave
Damn, I meant to include Brave and somehow it slipped away from me. Sorry!
♥ 1
"Online advertising certainly funds something. But is that something the open web? That looks rather debatable at this point."
Great article from @riptari, loved some of the details 😂
Quoting a tweet by @riptari ↗
♥ 6
Replying to @null and @dansereduick, @CMAgovUK
Right, I have been thinking about whether we could treat Google's parts as separate businesses and require the transfers to be clear. I'm not sure how that's argued or established, though.
Replying to @Pedigo_Chris
Please do!
Replying to @publictorsten and @acfou, @CMAgovUK, @johnnyryan
I think that may be the case but we need to research the whole thing more clearly. It also introduces some opacity as to which creatives run on your site, which is a problem.
♥ 1
Replying to @publictorsten and @acfou, @CMAgovUK, @johnnyryan
It is! But entrenching that isn't a good idea IMHO.
♥ 1
Replying to @JInterlandi and @NafeesaSyeed
The number of people who told me "I'm sorry I couldn't get much done with that whole thing." I don't think we do enough to counter that guilt. It's deep-rooted and toxic.
♥ 1
Replying to @null and @amyvdh
I think I just got lucky 😁 Funny how the vibe is similar to this one, beyond just the words.
One of the lessons of this pandemic is that governments often really, really suck at data, & don't care much about fixing it.
The quality of the French epidemic data has been abysmal, the official government dashboard was republishing the work of volunteers. Vaccine data? Same.
Quoting a tweet by @redouad ↗
♥ 16↺ 16
Here goes my 2021 reading log! I anticipate fewer books this year as I have started switching to taking more in-depth notes. But we'll see!
♥ 53↺ 8
If you're looking for my 2020 log, it's over there. Was I ever young, foolish, and innocent way back then!
♥ 5
"The Ethical Algorithm", by @mkearnsupenn and @Aaroth.
An excellent introduction to the field, focusing on solutions (on the assumption you know there are problems). Nontechnical, but usefully applicable in the trenches.
♥ 18↺ 1
Read the whole thread: 84 tweets, continued until 28 Dec 2021 →
Replying to @jwajsberg
Je ne dis pas que c'est identique, c'est juste pareillement nul 😁
Replying to @jwajsberg
Pareil, mais dans les deux cas ils n'ont pas fait d'open data, pour les données épidémiques le gouvernement a juste fait un dashboard avec des données récoltées par des volontaires qui se coltinaient les PDFs des ARS, tous différents...
There's a proposal from @k_johnsn in front of the AMP Advisory Committee to sunset AMP and AMP caches within two years, mainlining the good parts into Web standards. Discussed later today.
I'd love to hear feedback from the Web community on this!
github.com/ampproject/met…
♥ 40↺ 12
Replying to @johnbiggs
I can think of a bunch of use cases, maintaining a buy list and discovering new reads would certainly be interesting — though I would still want to have threads here to share with this gang.
holly.lol has a bunch of layout issues for me :(
Replying to @johnbiggs
I can't screenshot the whole thing because it scrolls way past the width of the window, but here's the right part of it! I also find it hard to read the light grey on purple.
Replying to @amarkeev and @tunetheweb, @k_johnsn
The idea is to be open minded about that. Different people have different things they like. Eg. I like declarative ad and tracker elements, even though I would not do them the AMP way.
♥ 1
Has anyone written a good argument to share to non-technical folks to convince them to drop WhatsApp and pick up Signal?
♥ 86↺ 20
Replying to @briankardell
LOL, yeah, I might need a little more ;-) That argument might work better from @BrobinBerjon.
♥ 2
Replying to @norsemelon
Yeah, I saw a few of those too. But not enough to tilt the balance!
The worst part is that WhatsApps looks like the worst parts of "Linux On The Desktop" cobbled together, so it's not like the UI is hard to beat.
♥ 1
A diagram you might want to spread around to move your friends off WhatsApp.
Quoting a tweet by @DavidNemer ↗
♥ 11↺ 7
Replying to @adambroach
A lot of those people *are* in Europe, and frankly if it's illegal that means they don't need to worry about it, right? *sigh*
♥ 2
Replying to @catthekin
Thanks, yeah, that's close to what I'm trying. Signal is even nicer on mobile, I find! I'm seeing little uptake with 2 and 3; 1 does help some. So far moved one group.
♥ 1
Replying to @yatil and @MoritzGiessmann
Yeah, this needs to work for people who think they have nothing to hide...
♥ 1
Replying to @clancynewyork and @morqon, @KateRoseBee, @Forbes, @UKZak
Thanks — yes, I think this is one of the best resources people have sent so far. The picture says it all, really!
♥ 2
Replying to @null and @amyvdh
I hadn't thought of the stickers. They really are cute!
Replying to @lolrepeatlol
It's true that it looks nicer! WhatsApp is shockingly ugly for something this big :)
♥ 2
Replying to @nataliabielova
I like the blackmail approach, I think I can get some groups to move that way — the hard part is going to be things like school parent chats and the such!
♥ 3
Replying to @mustach_io and @nataliabielova
I don't have an iPhone, but doesn't it fall back to SMS?
Replying to @fatemehx2
That's okay, I've convinced friends and family I'm useless at tech support anyway :)
Replying to @alextcone and @k_johnsn
Thanks! Even just knowing there's support in the broader community is helpful. The meeting went well, the AC seems to have very clear consensus that this is the way to go.
♥ 1
Replying to @alextcone and @k_johnsn
Ain't no shaming going on! Different AC: this is the AMP AC, which is an independent body set up by Google to show that AMP is an independent body with a mind of its own.
Well, it's doing just that!
♥ 1
This could also work in Europe, if only Ireland had a data protection authority to enforce the GDPR.
Quoting a tweet by @ashk4n ↗
♥ 5↺ 2
Replying to @stimpee
Hey, that's cool!
Replying to @stimpee
Ha, sure! Could be fun 😁
Replying to @MariamTadros87
Thanks!
Replying to @Chronotope
Imagine what it's like to speak French, where there is no word for "privacy".
♥ 3
Replying to @Chronotope
Yeah. It's usually translated as "respect for private life" which gets exhausting fast (and isn't entirely correct). There's an old word ("privauté") but it's not an exact match and it fell into disuses in the 17C.
♥ 2
Replying to @swodinsky and @clancynewyork, @anniesullie
That's not Verizon, it's a scam. My money's on a bad ad taking over.
♥ 8
I don't know if you can save your browsing session, but your best bet at preventing the same happening again is probably to switch to Brave or similar. Chrome isn't the best for protection.
♥ 2
Replying to @clancynewyork and @alextcone, @swodinsky, @anniesullie, @gillisusa
Sorry, I didn't mean to imply you're a rube! I answered the question I thought I saw, didn't mean anything by it. This isn't the browser taken over, it's the site, but the remedy is with whoever is serving that ad.
Replying to @clancynewyork and @alextcone, @swodinsky, @anniesullie, @gillisusa
Not so much the ad server as the ad itself. Basically, code in the ad can go "if (thingsLookSafe) then doBadThing()". That means that when the ad server runs tests on it, it looks safe and well-behaved, but when it gets a shot it can act badly.
♥ 1
So more like a customer at the bar was acting all nice and friendly, buying rounds, waiting for a drunk to exit and pounce.
♥ 1
Replying to @clancynewyork and @steveglista, @swodinsky, @anniesullie
Right, but this is where the bad ad sent you to, it's different from the bad ad itself. The odds are that site will be either operated by a company you can't easily sue, or it will have been hijacked itself.
To find the bad ad is harder, unfortunately.
♥ 1
Replying to @alextcone and @clancynewyork, @swodinsky, @anniesullie, @gillisusa, @Clean_io
Indeed — that's why we need better creative formats that can get us closer to static analysis and have a built-in audit trail.
♥ 3
Replying to @clancynewyork and @steveglista, @swodinsky, @anniesullie
You're not thick at all — you're working against a malicious party that's using one of the most opaque and complex ecosystems ever created to cover its tracks and try to trick you. This is hard!
♥ 1
To track down the ad, you would have to:
• Go back to Crimson
• Somehow get the same ad served
• Look at the source to find the ad call and the serving stack
• Use that to report to the Crimson, so they can report to their providers, who can block the seat.
♥ 1
This is hard because you'd have to dig into dev tools likely a few iframes deep. If the perpetrators are half smart, you'll never get the ad again because they are probably targeting away from you now or dropped a cookie so the malicious code doesn't trigger twice.
♥ 2
Replying to @alextcone and @clancynewyork, @swodinsky, @anniesullie, @gillisusa, @Clean_io
I think that idea is still alive — we were talking about using bundles for that just yesterday: github.com/WICG/turtledov…
It would be very interesting to see the audit trail get included at the CBOR level!
♥ 5↺ 1
Replying to @clancynewyork and @steveglista, @swodinsky, @anniesullie
Hahaha — the odds I can get the same ad are unfortunately very low. I just tried a few times and no dice. That's part of how they get away with it.
And OMG are the Crimson's programmatic ads bad. I even got a new variant on the buttflap! @swodinsky would be proud.
♥ 2
Replying to @alextcone and @clancynewyork, @swodinsky, @anniesullie, @gillisusa, @Clean_io
This is not SXG bundles, what we're looking at here is single-origin bundles. This does not reproduce AMP, but it does offer strong guarantees at the client level that the creative cannot have any content beyond what's right there.
♥ 3
Replying to @doriantaylor and @Chronotope, @clancynewyork, @steveglista, @swodinsky, @anniesullie
There's no foolproof way but they only need "good enough". See github.com/sindresorhus/d… for instance.
♥ 5
Replying to @alextcone and @clancynewyork, @swodinsky, @anniesullie, @gillisusa, @Clean_io
That's what we have standard orgs for :-D
♥ 3
Replying to @alextcone and @clancynewyork, @swodinsky, @anniesullie, @gillisusa, @Clean_io
Whoa, arrogant much, man? It's a lot less dysfunctional than the IAB and produces much higher quality tech!
♥ 3
Replying to @alextcone and @clancynewyork, @swodinsky, @anniesullie, @gillisusa, @Clean_io
I would say that "the ship has sailed on it being anything other than a quarreling cartel" is a pretty strong statement to be making from someone who works for another standards org.
All SDOs suck and the W3C has issues, but that's dangerously reductive.
♥ 1
Regarding the fact that it would be good to have standards for bits of adtech that can work not just on the web, I agree but W3C is the only org working on advertising in which publishers can have a voice, so I'll make do with what we have!
♥ 1
Replying to @alextcone and @clancynewyork, @swodinsky, @anniesullie, @gillisusa, @Clean_io
Alex, I jumped at your statement *because* I don't assume the worst of you. I found it to be a surprisingly insulting position for you to take. It's also not factually accurate.
♥ 1
Replying to @alextcone and @clancynewyork, @swodinsky, @anniesullie, @gillisusa, @Clean_io
Agreed, agreed.
♥ 3
One thing that's cool in the US is how easy it is to pick a mobile phone plan. They only have three operators, they're all ridiculously overpriced, and only one has a website that doesn't crash when you try to order a SIM. Much easier than that free market economy nonsense.
♥ 27↺ 5
Replying to @mikarv
I mean competition is, like, so exhausting! And look at how well tech is working without it.
♥ 2
Replying to @mikarv
I know, when I moved here, it was cheaper to keep my Free line and roam than to get something local. I only eventually switched because people didn't j understand the phone number.
♥ 1
Replying to @justinph and @_alastair
Monopolies is why they can keep making switching suck and pay for advertising :)
If you don't read French, you might not know about @opquast's outstanding Web Quality Assurance checklist. But — lucky you! — it's now been translated into English. You can read @ElieSl introduce it here:
opquast.com/en/opquast-who…
♥ 9↺ 1
Replying to @tbernard1979
Rien de bien génial franchement.
Replying to @jarroddicker and @JayCoDon, @Chronotope
Yeah I have thoughts: burn it with fucking fire then salt the earth with plutonium waste 😘
A core tenet of the Web is trust, and the fundamental idea behind capturing and using email in this way for ad targeting is a violation of trust.
♥ 11↺ 2
It's not going to be "please give us your email address so we can give you a better ad experience" (because users know better), it'll be "here's a regiwall for no reason" and users will just use SSO to get past it because they're used to it, without really consenting.
♥ 3
Then the site will turn around and betray the user's trust by sharing it with poor protection (hashing which can be reversed, if only because of the scale).
Eventually, this will destroy our ability to rely on email for legitimate uses.
♥ 7
I'm thinking about a proposal to get rid of email entirely on the Web, it's a bit of a long shot but we need to get there. Cross-context tracking is dead, period.
♥ 9
Replying to @JayCoDon and @jarroddicker, @Chronotope
Ceterum autem censeo Carthaginem esse delendam!
♥ 2
Replying to @DinaSrinivasan
In the app? I can see it touching cookies in its Web version, but I'd be quite interested in understanding where it sees cookies in apps. I don't think I've seen it having an in-app browser, maybe to load URL previews? But does it do that as the user?
♥ 1
Replying to @kfranasz and @DinaSrinivasan
Maybe! Also setting up a site to see what it does to fetch previews (I think it does).
♥ 1
I don't know if it's ageing or hitting the VP level, but I miss the time when I got spam that wasn't just people inviting me to webinars.
♥ 13↺ 1
Replying to @gabrielazanfir
To attend! Usually about things that are completely unrelated to my job.
♥ 2
Replying to @jlowcock
Shit, I was hoping I still had growth ahead of me, I've still got a good 25 years of work left!
♥ 2
Replying to @gabrielazanfir
Sadly none of the spam ones are even remotely tempting! But yeah, even interesting conferences I feel like skipping these days, it's just not the same.
♥ 2
Replying to @hyacinthsystem and @beccadottex, @swodinsky
If the switch happened with no noticeable change that would be very strong evidence that someone has a knob they can fiddle with to tune prices artificially...
♥ 2
Replying to @swodinsky and @DinaSrinivasan
Is it not still only a preprint? It's still time to update it 😁
♥ 2
Replying to @csarven and @Miel_vds
I didn't actually know we were involved in that!
Governmental agencies are standardising on ad blockers because adtech is too much of a malware vector (it also makes it easy to spy on users).
How long before this is expected of CISOs in business too? How about we fix this before it's too late?
cyberscoop.com/ad-blockers-se…
♥ 101↺ 45
Actually, I'm pretty sure that knowing JavaScript has become the *only way* to check the PR slop of a massive international tech conglomerate.
Quoting a tweet by @swodinsky ↗
♥ 6↺ 1
Replying to @johnwilander
Yes, and just think about the intelligence you can gather from the bidstream.
♥ 2
Replying to @kenarchersf
There's a whole number of ways for advertising to function, including for smaller publishers, without as insecure practices nor anywhere near as much privacy violation.
♥ 4↺ 1
Replying to @jasonkarns and @AMPetry, @LeonieWatson
In fairness, a lot of ad blockers are pretty dumb and brutal. I worked on a CMS that loaded content from the frontend based on the URL, and they would just block any page that had "advertising" in the path.
♥ 1
Replying to @TheAnnaGat
French breakfast is café au lait and pastry that's literally shaped after a Muslim crescent, both of which became European after the Turks lay siege to Vienna.
♥ 2
Replying to @fgeorges and @jlowcock
Hahaha, it's long! I've been working 25 years and the idea I have 25 more to go makes me wonder how I'll keep it interesting!
♥ 1
Replying to @montezumachavez and @hackylawyER, @Google
Probably, but...
tweets.berjon.com/1347633616533278720 (cc @ProtonMail)
♥ 2
It's easy to read this as just some technical cheating over ads from Google and Facebook in collusion. But this deal alone made sure hundreds of millions of dollars earned by publishers were kept between those two companies. 🧶
nyti.ms/3imRzwt
♥ 61↺ 35
This can be measured directly in journalism jobs. There are stories like the #MeToo ones or the Trump Taxes that none of us know about as a direct result of this "Jedi Blue" deal.
Let that sink in.
♥ 8↺ 1
It's also not isolated, it's a system. Most of what Facebook asked from Google in that deal was that Google wouldn't get to cheat with them as they do with others (hide the prices, reuse the data, keep identity to itself).
♥ 7
And what do publishers do if their ad revenue is held artificially low? They increase the volume and invasiveness of ads. In turn this makes AMP attractive, which furthers Google's strategy of running the Web the way it runs YouTube, like a toxic strip mall.
♥ 10↺ 1
The silver lining in this is that it is pretty clear that they cheat because they couldn't compete on a level playing field. With just a bit of levelling we can both get democracy back and enjoy watching them reinvent themselves as service companies.
♥ 13↺ 2
Replying to @Tralallama
It's hard to tell, there's a lot going on in parallel. I think it's likely but it might take a little while longer.
♥ 1
Replying to @null and @PaulDavisThe1st, @johnnypsmith3, @ZachWeiner
Sorry, but this is simply not the case. Some portion of the loss of ad revenue in media was likely justified (prices were very high) but the system that Google and FB have put together today has nothing to do with economic efficiency and everything to do with monopoly power.
Replying to @null and @PaulDavisThe1st, @johnnypsmith3, @ZachWeiner
They exact rent by forcing media to do work to exist on their platforms (see AMP). Also, we do buy a lot of advertising, to drive subscriptions as a substitute, and some do it to drive traffic, hoping to get something go viral.
The way I understand "Jedi Blue" is that it's the work of men (I assume men) whose sense of ethics is entirely summarised by the fact that, if their project has a codename from Star Wars, then they must be the good guys. nyti.ms/3imRzwt
♥ 22↺ 2
Replying to @flamsmark
I completely agree, I don't think that we can save the old web, but the question is whether we can build new things with similar properties, at meaningful scale. I tried to explain IRC to people the other day, but the idea of nonproprietary chat is weird. Same for the Web?
♥ 6
Replying to @tim_libert
It's weird, I thought it worked on ISIS because brown skin people hate freedom.
But let's give every user a unified ID to make them easier to target (by malware campaigns)!
Quoting a tweet by @SwiftOnSecurity ↗
♥ 2↺ 1
All recommendation systems suffer from shit in/shit out problems. But recommendation systems that also learn from the engagement they generate have shit in/shit out in a feedback loop.
It's Ourobouros meets human centipede. This makes me see the name "YouTube" in a new light.
♥ 25↺ 3
Replying to @RichFelker
For the Web to be financially sustainable.
Replying to @plscallmekimi and @flamsmark
Because most people on the net used to use it and now it's a small minority.
Replying to @plscallmekimi and @flamsmark
Making something easy to use isn't bad, it gives you time to use fluency on other things. If you want to talk to someone, having to think of which server they're on, what their nick is, and run a screen session so as not to lose messages is hard for no good reason.
♥ 4
Replying to @opendna and @kyhwana
I think there are ways to implement ads better than adtech so far has. Watch this space.
♥ 1
Replying to @cepcam
Absolument. J'ai fini un peu sur une sale blague, mais ce à quoi je songeais ai départ est l'effet sociétal prévisible. Si tu pars d'agents plus ou moins sains, et que tu les expose à ce genre de système, il semble logique que ça les rende un peu dingues.
Mais à court terme et selon les spécifications le système marche très bien 😁
Replying to @BuggeErik
Absolutely!
Replying to @null and @nhoizey
Twitter automatically uses AMP, so if you share from Twitter, as I did, it'll do that. We should get them to change that!
Replying to @null and @SimonDeDeo
We've all been too Reaganised the world around to even consider reaching for the kind of state intervention into private production capacity this would need as being in the repertoire of options.
♥ 1
Replying to @null and @nhoizey
Yeah, it's not great! But I don't even think that they use the preloading or anything. I should find someone to talk to there. They've often been better at doing the right thing.
Replying to @tim_libert
Hahaha 😂
Replying to @null and @sivavaid, @cwellmon
There are also some i interesting notes on the topic in fivebooks.com/best-books/pol…
♥ 1
25 years ago today nytimes.com went live!
Gotta love that crazy 1996 feeling when people would launch a major news website *on a Friday evening*.
♥ 139↺ 32
Replying to @daveyalba
I was already feeling old because that's the year I began working, but now you've gone ahead and called me "respectable," that's... that hurts, Davey! 😄
♥ 2
Replying to @daveyalba
Oh, thank you :) I wasn't respectable on purpose, I promise to be more careful in future.
♥ 1
GPC makes it into The Times, even including a link to the @w3c draft!
nytimes.com/2021/01/19/tec…
♥ 17↺ 5
Replying to @anglobibliofile and @Slate
Honestly, that's kind of civilised, I could see that model having a future :)
♥ 4
Has anyone here mustered the patience to watch the second episode of WandaVision? Is it worth trying? I wanted to like it but when I saw ep2 kicked off with the same tedium that filled ep1 I just couldn't...
♥ 2↺ 1
Replying to @DannyPage
Yeah, maybe when there are more episodes I'll fast forward to where something happens!
Replying to @steveworkman
That doesn't do much for me I think 😕
Replying to @gizmomathboy
Exactly! At some point I thought "I like how they're not afraid to have this thing last for a whole half episode!" But that got old...
A beautiful intrication of irony: the very first citation of this book on hypertext is Ted Nelson on the failings of the Web, and the citation call-out doesn't resolve. ♥️🔥
♥ 2
Wow this is huge. I love seeing some innovation in the browser like this, outside of privacy it's been in short supply.
Quoting a tweet by @dietrich ↗
♥ 13↺ 3
Replying to @DanielSolove and @daniellecitron
It's particularly low considering the revenues of some companies that don't seem particularly keen to comply...
♥ 1
Replying to @mnot
I'd be interested to hear more about the issues you see there.
"Google’s core arguments on the privacy and power front are actually really similar in one specific way: They only ring true if you don’t actually question Google’s claims."
Quoting a tweet by @swodinsky ↗
♥ 13↺ 2
Replying to @swodinsky
Oh no you're not. That blog post felt like they felt their arguments were so weak it had to be taken out with the trash in the middle of a long weekend.
♥ 1
Replying to @JGlogovsky and @slayser8
Well, if you *love* very old school TV and somehow Bewitched was one of your favourite shows, you might like it! Otherwise there's Black Widow coming, as well as Loki.
♥ 2
Replying to @swodinsky
I mean... he wrote that publishers find AMP wonderful the same week that the AMP Advisory Council resolved that AMP should be sunset in two years.
♥ 5
Replying to @ocdtrekkie
Please tell me when it starts to make sense! 😂
Nozick cracks me up.
♥ 5↺ 1
Replying to @hober
This is an endnote in just that :) I've only read chapter 1 but so far it's amazing.
♥ 1
Replying to @odemeulder
Walled gardens were always there!
Replying to @espadrine and @mnot
I thought I saw them mention IP tracking, maybe it was somewhere else in their docs.
There certainly are issues with IPFS, but I was wondering if Mark had some specific problems in mind with IPFS and HTTP being available in the same client.
♥ 1
Replying to @dmitriid and @BenDelarre, @Rich_Harris, @slightlylate, @matthewcp, @mikesherov, @brianleroux, @youngElPaso, @swyx
More generally, it's not as if we're short on cowpaths to pave. Just cracking open web dev projects provides plenty. Like, how many declarative routing implementations do we need before we get a standard? Complex forms?
♥ 3
I mean — has anyone looked at the innards of a rich text editor recently?
♥ 2
Replying to @dmitriid and @BenDelarre, @Rich_Harris, @slightlylate, @matthewcp, @mikesherov, @brianleroux, @youngElPaso, @swyx
For CSS, I've seen a whole lot of cases that could use support for simpler code management and reuse, that's a lot of what preprocessors solve. Extensibility? Meh.
♥ 1
Replying to @montezumachavez and @profcarroll, @airavn, @finnmyrstad, @ariezrawaldman
Actually, I suspect they actually mean less relevant in the sense that they'll target you only with ads you find terrible. Every ad I've had since making that choice has been some sportsy thing, it's like they're trolling me.
Replying to @profcarroll and @montezumachavez, @airavn, @finnmyrstad, @ariezrawaldman
Well, I'd like to think The Times has figured *some* of this out 😉 Seriously, I'd kill to understand why platforms with the content and behaviour that Twitter or Facebook have can't target effectively without tracking. Twitter shouldn't need to do this.
♥ 2
Replying to @edwinhayward and @uk_domain_names, @aral
They might be A/B testing their social postings.
Replying to @_alastair
I remember Sliders! But it's true that no one gets the reference. I often think of the one in which Napoleon decided to keep French Louisiana and make peace with England, leading to a French USA full of people on bikes carrying baguettes and wine, and giggle.
♥ 2
"You know who manages by metrics? Big companies like Google, Amazon and LinkedIn.
And what do they have in common?
Their core product got notably worse over time."
Some really good, hard truth from @jakobgreenfeld over at jakobgreenfeld.com/metrics.
♥ 53↺ 14
Decisions are hard and people rarely want to actually be responsible for anything (they just want the trimmings of being "in charge"). The easiest cop out is to make everything "data-driven" (instead of data-informed): that way you can avoid a whole lot of responsibility.
♥ 12↺ 3
A further problem is that your users will be more legible when they have less autonomy and are made more uniform. So data-driven product development has a vested interest in disempowering users.
People don't decide to disempower users — but features that do will measure better.
♥ 5
Whenever I hear a company boasting that it A/B tested 200,000 shades of blue and shapes of a button all I hear is that they must be really shit at making products. That's usually confirmed with a quick look.
♥ 8
TL;DR: feed statistics to your designers, not the other way around.
♥ 11↺ 1
I forget to say: heard via @podopie!
♥ 1
Replying to @__apf__ and @jakobgreenfeld
I think it's an industry-wide problem (for a broad understanding of "industry") and I agree it needs some good thinking and talking.
I got the impression Google was given as an example of a well-known product that has been headed that way more than of laser focus?
♥ 1
Replying to @__apf__ and @jakobgreenfeld
Ah, I see what you meant by laser focus. I don't read the post as stating that you need hard metrics from top to bottom and across the whole org.
The eg. of SERP, which has become markedly painful over time, feels like it could happen with just a few people looking at a metric?
In that sense it feels to me very similar to LinkedIn (you want an easy professional address book, you get stabbed in the face with prompts to install an app) or Netflix (you want to watch something but you can't even find your own watchlist again because recommendations).
♥ 1
Replying to @__apf__ and @jakobgreenfeld
On this, I am not certain that I fully agree. I think it depends on product philosophies. Products that are tools, that enable people to increase their agency and do more themselves, can scale to 1bn users because each user can make their own choices.
Products that are more passive-oriented, and that try to do things *for* users and in their stead, will indeed have a harder time scaling because they can only work on some assumptions of legibility and uniformity that at best aren't true, at worst are dangerous.
Replying to @benadida and @jakobgreenfeld
He literally says "That doesn’t mean that I’ll not use a fitness tracker, make regular blood tests, or check my bank account. But I certainly won’t allow them to dictate my decisions."
♥ 2
Replying to @swodinsky and @randfish, @dohertyjf
Ha, "adtech expert" — what can I have done to justify such ire, @swodinsky?
Jokes aside, this is… complicated. I'm still mulling over what a good approach would be.
If Google still ran a "turnstile" SERP that mostly directed traffic elsewhere, I don't think they should pay.
♥ 8
Of course, that's what they say they do. But it's pretty obvious to everyone else that that's BS. The majority of the links in SERP point back to Google. They use knowledge others have invested to build for snippets, voice, etc.
♥ 5
There's also AMP, that benefits Google at a high cost to publishers in terms of CPMs, conversions, implementation. They keep saying that they don't derive much money from news and yet they keep coming up with ways to force news content to be hosted on their own platform…
♥ 7
I mean AMP, News, Showcase, Discover, SXG… they'll basically do anything other than let publishers host their own content.
At some point, if they keep pushing in that direction, they should expect to pay for it. But I have strong misgivings about heading there.
♥ 9
Eventually in that world we end up with Google running Web content (at least news) in the same way they run video with YouTube — aka like a train wreck made of dumpster fires.
♥ 10
My conspiracy theory is that, in its negotiations with @Adlc_, @acccgovau, etc. Google is playing hard to get to give them the impression that they're resisting, but it's actually what they want: host the Web as its own walled garden and control monetisation.
♥ 24↺ 1
Replying to @randfish
They always tell me I'm better at evil than they are. But if I'm wrong then I'd like them to stop coming with schemes to host the entire Web :)
Replying to @dohertyjf and @swodinsky, @randfish
If you replace "left behind" by "lose most of your traffic and therefore source of revenue" then yes!
♥ 2
If there isn't a GAN to do this I think we should all agree that machine learning is useless.
Quoting a tweet by @donohoe ↗
♥ 4
Replying to @JackBSmoke and @swodinsky, @randfish, @dohertyjf, @Adlc_, @acccgovau
I think Facebook is already AOL, they just haven't realised it yet. In my view, Google is more like Microsoft: smart techs, ruthless business, but their products succeed mostly from market leverage rather than hearing feedback so they're increasingly all shit.
♥ 1↺ 1
Replying to @mnot and @swodinsky, @randfish, @dohertyjf, @Adlc_, @acccgovau
Yes, though I'm not sure that the @CMAgovUK will necessarily act in that direction, at this stage they're just looking.
Replying to @tim_libert and @swodinsky, @randfish, @dohertyjf
It's the same company 😁
Replying to @mnot and @swodinsky, @randfish, @dohertyjf, @Adlc_, @acccgovau, @CMAgovUK
Do share 😁
♥ 1
Replying to @tim_libert and @swodinsky
That's fair. Honestly, I think it's more intricated than that. YT results show up in SERP video, there is overlap in the ad stack. I think that you see a difference because YT is a social media so has a different kind of impact. But had Plus worked it would be the same IMHO.
♥ 1
Replying to @manjusrii
I agree with you that neither of those two points are credible much. (What "open Internet"?) I'm more worried that it builds an institutional structure that actually maintains the platform power. Curious what you think!
Replying to an earlier tweet of mine and @swodinsky, @randfish, @dohertyjf
My conspiracy theory is that, in its negotiations with @Adlc_, @acccgovau, etc. Google is playing hard to get to give them the impression that they're resisting, but it's actually what they want: host the Web as its own walled garden and control monetisation.♥ 24↺ 1
Replying to @baekdal
When it's that simple, I agree. I think the problem is that the platforms, especially Google, are trying to have it both way: claiming to be a turnstile but also pushing to carry the content. They've muddied this a lot. Some notes in that short thread:
Replying to @swodinsky and @randfish, @dohertyjf
Ha, "adtech expert" — what can I have done to justify such ire, @swodinsky? Jokes aside, this is… complicated. I'm still mulling over what a good approach would be. If Google still ran a "turnstile" SERP that mostly directed traffic elsewhere, I don't think they should pay.♥ 8
↺ 1
I wish someone would organise a conference of "people who use Seeing Like A State to understand tech." It would be amazing.
Quoting a tweet by @math_rachel ↗
♥ 34↺ 5
Jim makes a good point in this good thread that the EU should take @ICOnews's performance into account in making an adequacy decision, and as things stand it's not looking great.
Quoting a tweet by @jimkillock ↗
↺ 1
Replying to @mnot
If it were a topic on which I had a better network I absolutely would, but most folks I can think of are just a little bit too far removed for me to barge in and have them talk.
♥ 1
Replying to @slayser8
I KNOW RIGHT I'M SO EXCITED THIS CAN'T NOT HAPPEN
Replying to @slayser8
OH NO DON'T DO THAT STEPH NOW I'M *EVEN MORE EXCITED* AND I CAN'T ALL CAPS MORE
Replying to @null and @ericnull, @joejerome, @Forbes
The worst part is that there's a small but non-negligible chance that someone tried to make this work by embedding the AdChoices framework. It's absolutely the wrong way to do this, but it would take a long processing time unintentionally.
Replying to @ns_ahmed
You and so many of us too 😁
Replying to @TzviyaSiegman and @mnot, @dauwhe
❤️ Well, I don't need more than the three of you there to want to attend that conference :)
♥ 2
The NYT EdBoard on terms of service is worth reading (also in the Sunday Review for those who do paper). I thought they did a great job pointing out the absurdity of this system, and yes, anything with some @hartzog in it is good! t.co/YIIODSoS4O
♥ 19↺ 2
On connait finalement assez mal cette partie de l'oeuvre de Spinoza où il écrivait pour la rubrique "Dicton du jour" du Dauphiné Libéré.
♥ 6↺ 1
Replying to @LFena and @hartzog
I think we've all been finding out the hard way that transparency actually doesn't do that much, at least not on its own!
♥ 2
Replying to @null and @vekstra, @hartzog
I'd be curious to hear your opinion of ours! We did a lot of work to make our privacy notices readable, but the ToS didn't get the same treatment.
♥ 1
I didn't know this was coming but I'm happy to see it as if I'd been waiting.
Quoting a tweet by @montezumachavez ↗
♥ 1
Replying to @mario_gug and @montezumachavez, @Tinojaramillo, @PiracyByDesign
That was my thought on this piece: this is a form of dark pattern. You promise transparency and choice, therefore agency, but you build it in such a way that it can't be exercised.
♥ 2↺ 1
Replying to @daniellecitron
It's not gigantic in absolute terms, at least compared to the disappointing ones the platforms have received, and I know Grindr has a bad history, but 10% of global turnover is still pretty steep?
♥ 5↺ 1
Replying to @daniellecitron
Oh sure, I mean, Grindr deserves a number of behavioural remedies — in a sense, no fine is enough! But looking at the bright side of this decision: if DPAs start to impose some real fines (and I think this is one), then we might get us some real GDPR! :)
♥ 6↺ 1
Replying to @jarroddicker and @JayCoDon, @Chronotope
Yup, good points there. There's value in cohorts, but that probably shouldn't be the primary reform we get from this current process.
♥ 1
Replying to @daniellecitron and @nytimes
Thank you Danielle, I've passed it on to the team. Thanks to you too, our work depends heavily on reading the kind of smart things you write! 🤗
♥ 1
Replying to @Chronotope and @JayCoDon, @jarroddicker
I think we should do more than assume good faith, we should put teeth into it. Chrome already backs up full history without any E2E — so the data is there for using anyway. We need a legal regime in which using a browser for self-dealing is illegal.
♥ 5
Replying to @JayCoDon and @Chronotope, @jarroddicker
I think we collectively need to recognise that this isn't going to work unless the browser is trustworthy — and that includes Google. I believe there is a shared interest in making that part work.
♥ 1
Replying to @JayCoDon and @Chronotope, @jarroddicker
Also, I disagree that there is little that publishers can do. The only reason publishers have such little leverage is because they don't show up and act as if they had no voice.
That's entirely untrue, but it's self-fulfilling.
♥ 1
Replying to @pbannist and @Chronotope, @JayCoDon, @jarroddicker
I don't see how that would be impossible, it could be happening right now! Chrome has way too strong ties to Google identity. Just looking at their SSO sideways logs you into Chrome. It's dark patterns all the way down.
♥ 2
Then when you're logged, it will prompt you to sync at every occasion. Breath on an extension or some preference and you'll get prompted. The prompt doesn't tell you much. It has a big okay button and it's not very obvious how to dismiss it. Again, dark patterns.
♥ 1
If you don't then go set up a passphrase for your sync, your data — including history — is synced to Google in cleartext. For some strange reason there is no constant badgering prompt to get you to set up a passphrase.
I'm willing to bet the overwhelming majority sync in clear.
♥ 1
Now, is this malicious? Maybe, maybe not. They have a good security team, but they never treat Google as part of the threat model.
Either way — they already have the data and can use it easily.
♥ 1
Replying to @pbannist and @Chronotope, @JayCoDon, @jarroddicker
What would that give you? The data is on Google servers. The identity is known. GAM can act on the full history of Chrome users without transferring data.
♥ 1
Replying to @pbannist and @Chronotope, @JayCoDon, @jarroddicker
Oh sure, they don't. But if you want to assume malice (I'm not saying I do, but if) then you need something to explain why their targeting works. FLoC can make for a nice smokescreen.
♥ 1
Replying to @idonibrasco and @daniellecitron
This could change, Ido! (via @montezumachavez) noyb.eu/en/austrian-su…
♥ 2
Replying to @montezumachavez and @idonibrasco, @daniellecitron, @maxschrems, @TetsuwanAstro, @EU_EDPB
One can hope the EDPB might change its mind! Right now the most effective DPA in Europe is arguably @NOYBeu - it shouldn't be this way.
♥ 4
Replying to @travelingheidi and @nytimes
I wonder how much of the original has survived to this day!
♥ 1
Replying to @montezumachavez and @ariezrawaldman, @airavn, @finnmyrstad, @mario_gug
That's hard to argue against. My hope would be that steep fines can help render user-first business models more attractive.
♥ 5↺ 1
Replying to @travelingheidi and @nytimes
Haha, yes, I can't imagine how many times it must have been copied!
♥ 1
Have any of you started looking into the NYS Data Accountability and Transparency Act? (Yes, they made it spell DATA.)
It seems to be somewhat CCPA-like — it could become law by April!
nyassembly.gov/leg/?default_f…
♥ 3↺ 2
Replying to @privacypen and @montezumachavez, @joejerome
Is he? I can see people getting excited about NYPA, but this one looks pretty underwhelming, no?
♥ 1
Replying to @null and @koalie, @pes10k, @SZimmeck
I think you'd have to pay :) The revolution gotta put bread on the table.
♥ 2
Replying to @joejerome and @privacypen, @montezumachavez
Ah, I see. Same: I'm excited because I think my team can implement compliance inside of a day (pending closer analysis). Beyond that it's quite meh.
♥ 1
Fasnitch. nyti.ms/3ojGMoc
♥ 3↺ 1
Replying to @null and @funnymonkey
It's strange, whatever part of my brain does empathy isn't triggering.
• 40 million users
• A bunch of browsers & extensions
• Major publishers and the Wordpress folks
• The biggest CMPs
GPC is real, folks — let's keep pushing!
Quoting a tweet by @globalprivctrl ↗
♥ 15↺ 3
Welcome to the resistance, Axel Springer!
businessinsider.com/big-tech-priva…
♥ 16↺ 6
Replying to @coolharsh55
The spec is deliberately staying a step back from legal prescription because that's something to be done in partnership with regulators. There are ongoing conversations which I find encouraging, happy to talk with more people if you have specific ideas in mind!
♥ 1
But basically the logic is that the browser is conveying the expression of rights matching sole controllership of the first party as equivalent to preventing sale in CCPA. So if data is shared to 3P controllers under consent, then withdraw, if under LI (or PI), then object.
♥ 1
Replying to @null and @davidhje
Cookie banners are a tough problem because they're legally mandated but do nothing for privacy. I'm not saying it's the case here, but you can provide great privacy and still have to provide a banner.
♥ 1
GPC satisfies the legal requirement for a valid consumer request to opt out of the sale of their data says California AG Becerra.
🎤💥
Quoting a tweet by @AGBecerra ↗
♥ 11
Replying to @null and @davidhje
Oh yeah, no argument there!
Replying to @mnot and @nytimes
Thanks for the feedback, they made a significant change: "A locked padlock tells us that a website is more secure, while an unlocked one suggests that a website can be more susceptible to attack."
♥ 2
I'm no lawyer but clearly a core tenet of American law is that if you disclose a practice in a notice somewhere then pretty much anything is legal.
Quoting a tweet by @Sally_Hubbard ↗
♥ 6
By reading the above tweet you agree that I can stab you in the face pretty much whenever I want.
♥ 7
Replying to @bbennettesq
I was going to, but then I remembered I can just change the terms whenever by adding to the thread so I figured I can cross that bridge when I smell the pitchforks.
Replying to @julien51
I think it depends on what you do, no? I could be wrong! But this practice can apply either way, they didn't invent it.
♥ 1
Replying to @emilybell and @chrismoranuk, @rasmus_kleis
The scariest words on the Internet are "We're doing this to help the publishers."
♥ 11↺ 2
Replying to @jason_kint and @FBoversight, @daiwaka, @MikeIsaac
A lot of Facebook's modus has been to do blatantly bad things with the certainty that there will be no real consequences and they can weather the bad PR. So far, it's worked.
I wouldn't be surprised if they backed down from nothing and fought this all the way.
♥ 2
Replying to @jason_kint and @FBoversight, @daiwaka, @MikeIsaac
Oh yeah, I get that. I wouldn't do it. But I can see them doing it anyway, in the same way that I could see Trump suing for defamation!
♥ 4
The Times is really pushing the envelope on graphics. nytimes.com/2021/01/29/tec…
♥ 45↺ 4
The Gray Lady, you said?
♥ 4↺ 1
Privacy matters, because legibility is oppression.
It's long, but if you read nothing else this month, read @shoshanazuboff's latest piece.
nytimes.com/2021/01/29/opi…
♥ 22↺ 12
She has this right: "A democratic surveillance society is an existential and political impossibility. Make no mistake: This is the fight for the soul of our information civilization."
This is ours to fix. It's going to require more than tinkering with cookies.
♥ 5↺ 1
I notably enjoyed her point that antitrust is important, but would it have been the best approach if we had understood climate change at the time of the Standard Oil intervention?
We're in the same position: big tech being anticompetitive is the smaller issue.
♥ 3
If we are to solve the bigger challenges of our times, we need working institutions. We won't solve the climate emergency with the same naive technosolutionism that's been churning out dystopia in the Valley. We'll solve it with democracy.
♥ 3
We need the imagination to reclaim what big tech has taken. We need to invert information asymmetry. We need governance over the attention commons.
We are in a dark place, but there is still hope.
Welcome to the Third Decade.
♥ 8↺ 1
Replying to @espadrine
Hahaha yeah Bloomberg have done some amazing things over the years, notably in web brutalism.
♥ 1
Replying to @fatemehx2
For the longest time I had no idea who did that, too 😁