July 2022
134 tweets
Oh wow!
♥ 2
On the principle I don't disagree, but this could also be a differentiating factor. Not all ML has to be opaque! Either way, I just meant this as an example of how to get there, I definitely won't say it's easy but I don't think it's insanely complex either!
♥ 1
Fingers crossed I can make it...
♥ 3
I REGRET NOTHING
Ha good one.
Hahaha 😂 I'm sorry, I don't understand the distinction you're drawing? 😉
♥ 2
It's The Immortal King Rao by @vauhinivara .I'll report when I've read it!
♥ 1
Hahaha, no it's this one.
It's The Immortal King Rao by @vauhinivara .I'll report when I've read it!♥ 1
Everything you need to know about both the state of the US and the state of Google Search in one single screenshot.
Quoting a tweet by @rustypolished ↗
♥ 17↺ 9
To think that there I was, naively thinking that a bunch of overt fascists seeking to control every uterus and the fight for trans rights as human rights were, like, different things maybe?
I'm glad @nytopinion woke me from that slumber! This has Contributed To The Debate.
♥ 6↺ 1
Ah, but both sides are actually the same! This is such original thinking.
♥ 2
Thank you!
Thanks a lot!
That's a very good point and I completely agree! There's a lot more to be written on this topic, notably on collective issues. I touch a bit on some collective aspects in berjon.com/stewardship/ but not specifically on the loss of self-discovery for communities.
♥ 2↺ 1
Sure — I don't think I meant to imply that absolutely *everything* was in that situation, only that it's a dominant part of the Internet experience. And I don't think it's intrinsic to the technology, we can fix this.
It's been around for a while! I have often mused about things to do with it.
♥ 1
FBI as in Funny Bad Ideas.
♥ 5
Yeah, it's big and not entirely cheap to process. But I've been interested in running health-of-the-Web experiments on it.
♥ 2
Beautifully written and brilliantly moving from @mariafarrell. There are many ways in which this speaks to depression, too.
crookedtimber.org/2022/07/05/set…
♥ 8↺ 2
Thank you 😁
Deleted!
Bluetooth headphones were a mistake.
♥ 15
They always fail!
I know right! Bluetooth headphones have now replaced printers as the dominant form of family tech support.
That is an EXCELLENT question!
No — all of them!
♥ 1
I had that experience too! But not in this case.
♥ 1
Thank you!
♥ 1
I don't know, I still like wires better. They Just Work™
I know I've been harping on building bridges between legal regulation and architectural regulation (from technology and tech standards) for a while, but look! A W3C CG draft document referenced in real-world regulatory rulemaking!
We can make this work.
Quoting a tweet by @montezumachavez ↗
♥ 13
Japan, a country of 125 million inhabitants, had only one firearm-related death in all of 2021.
♥ 7
🎶 You say Elon, and I say goodbye! 🎶
nyti.ms/3uB0HVw
♥ 2
More generally, the level of Dad Tech Support compared to wired ones is bad. Kids forget to charge them, they don't know how to pair, etc.
♥ 1
Subtweeting?
♥ 1
Offloading privacy labour to people is a form of victim blaming. One thing I've been wondering about for a while is the degree to which today's dominant approaches to digital design encourage learned helplessness.
Quoting a tweet by @EvanSelinger ↗
♥ 55↺ 19
The Chilean constitutional process is fascinating and the proposed Constitution seems amazing. I hope it gets translated into English soon, and that they vote for it.
Quoting a tweet by @saaaauuull ↗
♥ 4↺ 1
This is all the more remarkable when contrasted with the US that is collapsing under the weight of cultish worship of its antiquated Constitution.
♥ 2↺ 1
I've just been looking at amplification as the flip side of gatekeeping: it's deciding that a given audience gets to see given information. There's the organic kind (people repeat it around them) and the editorial kind (via recommandation algorithm or direct human decision).
♥ 1
Any specific amplification can result in cascades (eg. with an epidiological model), but the two are distinct. Looked at this way the issues are about the concentration of gatekeeping power (a tiny number of companies makes these decisions at huge scales) and...
♥ 1
...there are feedback loops between those companies' editorial amplification and their design of affordances for organic amplification.
♥ 2
I would add: the issue is emphatically not about terrible content (though that is a part of it). Rather it is about how we collectively decide what information rises to prominence. Editorial monoculture is more worrying than bad content.
♥ 1
Essentially, since we are epistemically social (but convinced of the contrary), the effects from gatekeeping concentration (editorially) and industrialisation (organically) are likely negative for our collective problem-solving abilities.
♥ 1
Any reading recommendations about the characteristics that distinguish extractive from retentive infrastructure?
♥ 3
But isn't this psychology? I'm looking for, like, roads and stuff kind of infrastructure. I mean not just roads, but that kind.
Are you thinking of something different from perceptual hashing? en.m.wikipedia.org/wiki/Perceptua…
♥ 7
Yes, they have all kinds of problems. "What people perceive" isn't a solved question, even assuming that everyone perceives in the same way, which we don't.
♥ 5
Palm Beach County, though.
I'm no expert either, I've just used perceptual hashing to find existing files in non-adversarial situations. I share your intuition that finding very similar pics is almost certainly a lot easier than detecting CSAM.
♥ 1
What I don't know is how trivial it is for a motivated attacker (who's making money) to defeat similarity detection. I would *think* that if nothing else, hashing various blur/resamples could help triangulate proximity in adversarial cases but I could be dead wrong.
♥ 1
Indeed!
♥ 2
"The broadest understanding of domestication as control over reproduction might be applied not only to fire, plants, and animals but also to slaves, state subjects, and women in the patriarchal family."
James C. Scott doesn't fuck around.
♥ 12↺ 1
Café nuisette.
♥ 3
With good jokes.
♥ 2
Un jeu de mots gourmand!
?
It's a shame to see ADPPA so diluted that this would be a concern. The original draft, despite its warts, showed a much stronger understanding of privacy and of what digital markets need.
♥ 2
Generally, it defaulted rather strongly to keeping data with the first party. That's better for contextual integrity and much better for publishers' businesses. You can't share data with third parties and keep publishers sustainable, it's a policy choice.
♥ 1
Yeah, you don't have to switch browsers to use the DDG app blocker. I browse on Firefox too, but block app trackers with DDG. It's far from perfect overall, but it's a clear improvement.
♥ 1
I've been meaning to try /e/, have you given it a shot?
Yeah, tell me about it...
Not in Alaska!
♥ 1
Worst offender is Keynote there. It's just bad.
Nice!
♥ 3
Non, c'est de la vraie.
Pretty please!
♥ 1
Thanks for writing it! 😁
♥ 1
The same is true of texts from those decades about digital technology enabling bureaucratic domination over people.
The problem is not foresight, it's designing capture-resistant forms of global governance.
Quoting a tweet by @alicebell ↗
♥ 6↺ 2
In this era of democratic backsliding, there is little that's more worrying than Google's irresponsible data handling practices.
Quoting a tweet by @MindingPrivacy ↗
♥ 35↺ 22
To put it more clearly, both my French grandparents spent the entirety of WWII fighting Nazis. Had Google been a German company in the 1930s, they'd be dead.
It's worth honest consideration: do we trust the state of American democracy enough to be comfortable with it today?
♥ 25↺ 10
Jeni is right: if you're legislating on privacy and aren't addressing collective issues, you're decades behind the problem.
Quoting a tweet by @JeniT ↗
♥ 4↺ 4
And resisted, too! en.m.wikipedia.org/wiki/Ren%C3%A9…
♥ 4
One interesting aspect of @netflix's foray into advertising is that it's bringing battles over privacy to TV.
Right now, TV systems are pretty leaky, with connected TVs and the likes of Roku observing and reselling behaviour. That's incompatible with Netflix's interests.
♥ 44↺ 9
You can only have a large scale sustainable ad business as a first party if intermediaries aren't able to extract and remonetise your data. Today's Web is captured by intermediaries for precisely the reason that it's a poor privacy environment.
♥ 12↺ 2
Those mistakes are being fixed on the Web, which is why you're seeing rearguard anti-privacy activism from intermediaries.
I look forward to seeing Netflix enter the arena.
♥ 9↺ 1
I know, and it's not like Samsung TVs are any better.
Maybe, but those laws are hardly being enforced so they don't matter much here. Also, consent only improves privacy under specific conditions that I don't think are likely in this case.
♥ 1
I was just reading @OlufemiOTaiwo about how the convergence of interests between elite and oppressed can be used to speed up justice. Alignment of business and people interests is a powerful special case of that!
♥ 3
I know, but it's not like GDPR enforcement is stellar!
I'm a French guy who set up GDPR at NYT - I know 😉
But most companies still approach privacy as a compliance rather than a product concern, with risk management lawyers rather than strategists. The result is at best milquetoast and most often against their own interests.
♥ 1
In this case my guess is that Netflix isn't going to be that blind and inept. They need to block Roku, Samsung, etc. from leeching data (rather than just getting some consent for them) and they'll fight for it.
At least, that's my bet & what I would do!
I'm curious: is there empirical evidence indicating that CPOs make a difference?
♥ 2
I'm not surprised 😁 But the rest of the space is... different.
♥ 2
No, they should provide infrastructure as service providers.
Haha, exactly!
♥ 1
This cliché annoys me no end. It may have been The Expanse's worst idea, even if they made it tolerable by having most screens be cracked. t.co/Wn3A1GAaIE
♥ 5
I suspect that it's a combination of phones look boring but are hard to improve upon and it makes what's happening on screen a lot easier to film.
♥ 2
Some interesting considerations below.
"Are we really going to accept an Interface Of The Future that is less expressive than a sandwich?"
Quoting a tweet by @Coleran ↗
♥ 4↺ 3
Loved this, thanks.
♥ 1
The UX of TV systems is appalling.
If they want to have advertising as a business model, they can provide ad infrastructure to first parties with contextual integrity. Or ads in their own experience areas.
That's a very good point! The only exception might be the Metaverse, and it's from the behemoth that most needs to capture infrastructure, which unlike the others it failed to do.
Yes indeed! So much of today's "future" thinking from technologists seems to be about making Star Trek happen - which is to say a vision of tech in which technology miraculously does not create social power differentials (but no one ever tells you how that happens).
There's good work on making parts of Internet/Web infrastructure into commons! It's not the subject of cool futuristic videos, though.
♥ 2
Can we buy tickets?
♥ 4
I wonder if the differences you point at with Plural Democracy are really ideological or have more to do with implementation strategies, some of theirs being naive or at least simplistic? It's hard to fully separate values from strategies, but there's room for it.
♥ 1
Voting is to governance what consent is to privacy: there are cases in which it's the absolute best option, but it's also the heavy-handed solution that too many people naively reach for and when overused it will backfire.
♥ 3
The alt models you list seem to me like they *do* have governance, it's just governance that's (very appropriately) limited to just what is needed to make things work. I'm not sure that goes against Plural Dem values?
♥ 1
A lot of commons models described eg. by Ostrom or in GKC have little or no voting, sometimes despite the fact that they manage significant infrastructure!
The label "democracy" might be an issue due to association with voting. I might call my branch "Cambrian Governance" 😁
♥ 2
Still France!
♥ 1
In fairness, those folks are looking into these issues with stuff like quadratic voting. But yeah.
♥ 3
I have the persistent impression that what you are looking for is essentially "governance done right" but I might just be ideologically biased!
♥ 1
Oh, I forgot to add: I don't know if you're familiar with the Aurora Apolito paper, but I think that it might be possible to infer answers to those questions that aren't in favour of voting from her use of IIT to describe the best social organisation.
♥ 1
In fairness, I don't think that things just got bad. I think that we're in genuinely novel territory for which we need innovative governance. That said, it's certainly the case that the big tech bureaucrats who have this power aren't working in ways that help at all.
I backed @benharnett's upcoming novel "The Happy Valley" and you could too!
kickstarter.com/projects/benja…
♥ 2
I look forward to reading it!
The @UnlockProtocol Litepaper has a very clear description of what it can do but also of the underlying philosophy. docs.unlock-protocol.com/litepaper/
♥ 9↺ 6
Forever.
She's looking for "a form of “collectivity” that everywhere locally max
imizes individual agency, while making collective emergent structures possible and
interesting." c4ss.org/content/52970
♥ 1
I don't agree with everything but I think that the idea of using IIT is very interesting. I suspect (but am still mulling it) that collectivity with local agency under those constraints rules out excessive use of voting.
I generally agree, but a few things to note:
* From an incrementalist view, it's better than what we have today.
* Not all token voting has to be plutocratic (though it's often a problem indeed).
* There are ways of using votes in limited fashion. Eg. for...
...contribution goods, where you contribute to be in but then most decisions are by consensus with voting mostly used for metagovernance (à la W3C).
I think that the form of DAOs that we've seen so far has mostly been rather naive and limited. I wouldn't assume that they're the endgame.
♥ 3↺ 1
I'm on the Internet every other day so I can't debate much but I think that equating all web3 work with (highly visible) dumbass cryptobros is a mistake. There's good work to be done, which is why looking at ideology matters.
♥ 3
I don't know, Aram, the past twenty five years haven't led me to believe that arguing over labels achieves much or prevents whitewashing. I'd rather just fix stuff so that bros don't matter.
♥ 5
Even infra stuff like Unlock, FileCoin, or the Graph?
More to say soon, I reckon m
No amount of decentralisation tech can address the user agent problem, that's why we need papers.ssrn.com/sol3/papers.cf…
♥ 3↺ 1
♥
♥ 3
most_interesting_man.gif: I don't whine often, but when I do, I do it on SSRN.
That paper is in bad need of serious editing. I've been letting it languish for a while, I hope to give it a full rewrite before the end of the year.
"Multiple brokers are likely hawking the same information, as half the world does not live in the United States, and half the world is not pregnant."
It's hard to tell if data brokerage is worse when the data is right or when it's wrong.
Quoting a tweet by @swodinsky ↗
♥ 22↺ 10
"Constitutions can thus be understood as an interface for a community’s shared values."
constitutions.metagov.org/article






