Replying to @LMSacasas
Why do you plan to ever stop?
♥ 1
March 2023
96 tweets
I'm running Standards, Governance, and DWeb Policy track at IPFS Thing (April 15th - 19th, Brussels).
Thoughts about standards+governance for a better Web? Ideas about joining public policy+tech to build sustainable digital commons?
Submit a talk at 2023.ipfs-thing.io/submit/!
♥ 6↺ 3
Replying to @DoraCrisan
I don't know, but I wasn't impressed with the PbD ISO draft when I saw it, so I'm not overly excited there.
Replying to @DoraCrisan
No always! When I checked it didn't even have a clear notion of personal data…
Notice & choice isn't just coercive, it's deceptive. We know *as an empirical fact* that people are deceived in their expectations and their understanding of privacy policies.
The FTC should treat processing data with this knowledge accordingly.
Quoting a tweet by @CenDemTech ↗
♥ 36↺ 17
Replying to @PrivacyMatters
Can it be done well? Absolutely. Can it be load-bearing to the point of defining a regime, at the level of complexity of today's data processing? I really don't see how.
♥ 3
I don't generally believe in propertarian approaches to data, but has anyone considered that people could have property rights over models produced based on their data (and their copyright)?
♥ 8↺ 2
Replying to @adambroach and @humanpropensity
Exactly. Individual data property is useless, doubly so over something that can't be priced. Voting rights (and dividends) over a collective entity might not be.
♥ 1
Replying to @MarkDonatelli
I'm aware of that, but it has too many downsides. Individuals jointly owning a collective thing however is more interesting.
♥ 1
Replying to @montezumachavez and @PrivacyMatters
I understand that that was the idea, and it probably worked in the 70s. But no notice can produce transparency or support informed consent beyond a given level of complexity or frequency.
♥ 3
Replying to @julien51
Are you thinking in the DAO space? Because most of what I've seen (but I certainly haven't seen everything) has been about rights over the data itself rather than derived products. I think the creation of a real artefact that's collective could shift the space.
Replying to @julien51
Property rights != ownership. They're the rules that apply to what people can do to resources. Ownership, in the strict sense of a transferable deed, when over a bag of bits, probably gives you nothing at all?
Replying to @Elibietti and @jennifercobbe, @montezumachavez, @PrivacyMatters
Note that my original point wasn't that we should get rid of notices. It was that, given widespread knowledge not only that they don't work but that they induce false beliefs in data subjects, relying solely on classic notices for data processing should be deemed deceptive.
♥ 3
On the question of transparency, I think that we should shift from transparency for people to transparency for research & investigation, ie. discoverable, machine-readable, etc. Simplifying collective action at scale has potential IMHO.
♥ 6
Replying to @internetdaniel and @Elibietti, @jennifercobbe, @montezumachavez, @PrivacyMatters
Hahaha, outstanding!
♥ 1
This is so amazing, I'm heaving.
We should have an "Innovation is my passion" variant for elected officials.
Quoting a tweet by @AntoniaRForster ↗
♥ 14↺ 1
Replying to @jennifercobbe and @Elibietti, @montezumachavez, @PrivacyMatters
/me stares pointedly at the DSA advertising metadata requirements.
♥ 3
Replying to @Elibietti and @jennifercobbe, @montezumachavez, @PrivacyMatters
Looking forward to all of these pieces! 💕
♥ 1
Replying to @julien51
Sure - that's where you can introduce a size threshold. It should be expensive (and safe) to be big
I was lucky enough to read this a few months back, you should read it too!
A powerful story with great emotional range, going places you didn't expect but seem unavoidable in retrospect, with a crisp, biting style.
Quoting a tweet by @mariafarrell ↗
♥ 9↺ 2
Replying to @juanof9
That's an interesting question, and I don't know! I wonder to what degree the copyright and personal data aspects connect, or if that's just superficial similarity.
♥ 1
Replying to @jchris
It's also captured infrastructure, which we need to not have. I suspect I know (and agree) with what you're going to write, but I look forward to reading it!
♥ 1
In which @msalganik tries to use ChatGPT to help with peer review, doesn't get any help from it, but finds cool things along the way. freedom-to-tinker.com/2023/03/08/can…
♥ 3
Replying to @mslkmp
Quoting a tweet by @LMSacasas ↗Interesting things might happen when you focus on making technology useful for people instead of bragging about the size of your parameters in the AI locker room. spreadprivacy.com/duckassist-lau…
♥ 13
Replying to @null and @MadronaCitizen
I mean LLMs are about nothing other than grabbing things without consent.
Replying to @CheckMyAdsHQ and @DCNorg, @jason_kint
I get the desire to isolate Fox, they deserve it many times over. But slamming allies over what amounts to a detail is — I'll be blunt — a really dumb way to go about it.
It's not like we're not already massively outgunned. I'm really not looking forward to movement bickering.
Replying to @nandoodles
No — by "dumb" I mean that there is no outcome from this that leaves Fox in a worse position and either you, DCN, or anyone who cares about the direction of ad funding in a better position.
QTing for trollbait isn't a strategy.
Replying to @KalOfTheRathi and @CheckMyAdsHQ, @DCNorg, @jason_kint, @catthekin
I know they do good work, I've been around a long time. That's why I'm taking the liberty to point out that this is a bad move with zero possibility of a good outcome. It's what friends do.
There is no step that the DCN can take in response to this that will have the slightest impact on ad buys. But irrespective of what the DCN does or doesn't do, this burns a lot of bridges. No one wants to collaborate with someone who'll throw you under the bus on a whim.
♥ 1
So yeah, net loss for @CheckMyAdsHQ, net loss for all of us figuring out how to fix this mess.
Replying to @null and @dansereduick
I'm still mad they cancelled it. It felt like they never marketed it? I sort of picked it up at random instead of their abysmal recommandations.
♥ 3
Replying to @daphnehk and @thepublicdomain
Thanks! I've enjoyed the other @GKCWorkshop books too, and folks in @metagov_project also have great resources.
♥ 2
Replying to @daphnehk and @rachel_grfn, @neil_chilson
You got there before I did, at least! But yes, I do look forward to newsrooms including GARM checkers in their CMSs so reporters can write to the rules. Great SLAS feedback loop there.
Ah — how it fills my little heart with joy to see advertisers defining responsible media!
♥ 1
Replying to @daphnehk and @rachel_grfn, @neil_chilson
What gives me hope is that the track record of brand safety tooling or this kind of initiative (like PRAM) is such that their odds of success are mercifully low.
It's almost as if power in ad systems lied with intermediaries instead of buyers or sellers. Weird.
♥ 2
Replying to @rachel_grfn and @daphnehk, @neil_chilson
Right, the fact that brand safety has become the tool to demonetise disinfo is wild. So far it has only succeeded in demonetising coverage from legit publishers when they cover some of those downer woke topics like "climate change."
♥ 1
I haven't seen many people ask why disinfo (and MFA sites in general) are kept profitable by managing to win bids for *targeted* campaigns even though their audiences don't produce actionable targeting information. They have no business without third-party tracking, basically.
Replying to @daphnehk and @rachel_grfn, @neil_chilson
Publishers don't talk about it openly much because they enjoy being beaten up, but some of it has made it out:
• businessinsider.com/publishers-adv…
• marketingbrew.com/stories/2022/0…
• adalytics.io/blog/tens-of-t…
♥ 2
Replying to @rachel_grfn and @daphnehk, @neil_chilson, @PlatGov
Well — social is a huge source of traffic and traffic is money, so it's pretty safe to assume that the impact is significant! It's unlikely that there's an analysis though because it's hard to tease out factors with those black boxes.
♥ 1
Two weeks from now will mark the tenth anniversary of the landmark RFC 6919 (datatracker.ietf.org/doc/html/rfc69…).
I think that we SHOULD CONSIDER celebrating.
♥ 9
Replying to @daphnehk
He basically invented ultramarathoning while researching data-driven approaches to carbs budgeting.
♥ 1
Replying to @_asmodai and @Esqueer_, @ZoeSchiffer
This cat meme is so good it's admissible in court.
♥ 1
Replying to @ineshipolito
So… more power to powers? :)
Replying to @OR13b and @ericlaw
We have a growing blocklist system which we're kicking off governance for in github.com/DDC-WG so that it's publicly accountable. We're also pushing towards subdomain gateways (safer).
♥ 1
Replying to @Web3Storage and @ericlaw
We also have ongoing work to add to that heavy-handedness! github.com/ipfs/in-web-br…
♥ 1
Replying to @msdixon
It doesn't stop there. From "Foucault" over your right shoulder, to "A Theory of Fun" a few books under the aristocrat, it's a whole family of moods.
♥ 1
Remember when people thought tech companies were on the side of democracy? LOL, fun times!
Quoting a tweet by @accountabletech ↗
♥ 14↺ 3
Replying to @edasfr
That they weren't evil, I don't think so, but that they were aligned with democracy was established policy, eg. of the Obama administration.
Replying to @edasfr
In what?
Replying to @gordonbrander
I really liked Pickering's overview, I thought it covered a lot of ground with just the right kind of detail.
♥ 1
Replying to @KrisShrishak
I don't think that either Google or the DPC have heard about purpose limitation :)
♥ 4↺ 1
When a company tells you it's shipping an experiment, the question to ask is: "who are the test subjects?"
Quoting a tweet by @mmitchell_ai ↗
♥ 10↺ 7
Replying to @random_walker
Do they show if you specifically dig not just on your profile page but in the replies tab? I've seen behaviour where they show only there, not even on the main profile.
♥ 1
Replying to @JulesPolonetsky
Put slightly differently: if you want to do psych pricing, you have to understand the psychology relevant to the purchase!
♥ 2
Replying to @random_walker
If I go to this I see four tweets instead of three, but still not seven. Anyway, I'm not surprised it's broken...
Quoting a tweet by @random_walker ↗
Replying to @KrisShrishak
And was your ethics team was fired over the experiment?
Replying to @ntnsndr
Lol, can you talk to yourself or still locked out?
Replying to @IPFS and @n0computer
I don't think that's b5?
Reading @superwuster's "Master Switch" and concluding, as is fashionable among some techbros, that the best we can hope for is a benevolent monopoly is like reading an IPCC report and deciding the best outcome is a climate change suntan. We have better options in reach.
♥ 5
IPFS Thing tickets are 45% off with code THING23! Get yours while it lasts!
2023.ipfs-thing.io
♥ 1↺ 1
I have a question but I'm unsure what to search for: is there a model that links the degree of species monoculture/diversity in an area to the likelihood that a predator or parasite will see explosive growth there and near-wipe the monoculture out?
♥ 1
The monoculture → "pest" link is often presented as evident (it probably is) but I'm guessing someone has formalised this and perhaps tied it to "degree" of monoculture (if that's a thing).
(I think this isn't Lotka–Volterra, but maybe next door?)
♥ 1
The whole *point* of LLMs is to enclose more of the commons. These companies see relations they don't intermediate as bugs and attack them with full force.
The riverbed won't be dry, just unfairly compensated to produce privatised content for the model.
Quoting a tweet by @collect_intel ↗
♥ 33↺ 14
Very interesting presentation on BLAKE3 & bao from @n0computer. I wonder if it would be useful to apply this in MPEG DASH?
youtube.com/watch?v=nk4nef…
♥ 2↺ 1
Has anyone compared levels of disinformation or Made-For-Advertising garbage sites in markets that have search (and social) competition versus those that don't? It's inordinately easier to SEO for one than for many, so search diversity should have an impact there.
♥ 5
One of the smartest people on these topics, hire her twice if you can!
Quoting a tweet by @hackylawyER ↗
♥ 4↺ 1
Space Karen asked to talk to the manager.
Elon Musk Tried to Meet With F.T.C. Chair About Twitter but Was Rebuffed nyti.ms/42N4r61
♥ 4↺ 1
If you like IPFS and you like specs, we now have a specs website! Get it while it's hot: specs.ipfs.tech.
Only a subset of the specs are published yet, we'll be iterating and pushing more out over the coming weeks and months!
It's not that I'm not interested in LLMs at a theoretical level, it's just that I can't help finding tech focused so blatantly on giving more power to already powerful people terminally boring.
♥ 24↺ 11
Replying to @null and @benharnett
So true. It's also very appropriate that Twitter decided not to notify me of this reply, presumably because you're one of those dangerous people who want, err, good things for people.
♥ 1
First iteration of my first IPFS spec today! 🪩🕺🏻
It's "IPFS Principles" and covers some architectural characteristics that are essential to the design of IPFS and to the IPFS ecosystem.
specs.ipfs.tech/architecture/p…
♥ 8↺ 3
I'm very thankful to be working on tech architecture with a caring and considerate community — it's been a delight. 💕
♥ 1
IPFS Implementations: It’s Definitely A Thing🌹💮
I wrote about the fast-growing ecosystem of IPFS implementations and how diverse they are in approaching different use cases, about and what makes IPFS, well, IPFS.
Get it hot off the IPFS blog's presses!
blog.ipfs.tech/2023-03-implem…
♥ 1
This thread is damning beyond belief, you should read all of it.
One possible read is that culture warfare is so far out of hand in the US that *even the richest elite is harming itself*. This is collapse territory.
Quoting a tweet by @jburnmurdoch ↗
♥ 2
Replying to @julien51
Thank you! If you found something meaningful in it that's a good measure of success.
Re federated: I think they can be part of the solution, but if that's all you have then yes, you'll get feudalism of some sort.
Basically, if you want various forms of cooperative governance, you quickly have scale problems. You can't have a coop with billions of people, and even at smaller scales you quickly hit the point where voting is the only way to aggregate preferences, which isn't great.
A way out is to have lots of small-enough communities, and they can use various options from CommunityRule. And then those get to federate.
I'm handwaving because I don't think we have the details. But yeah: federated benevolent dictatorships is a dumb way!
♥ 1
Replying to @julien51
I don't think that wide-scale voting is necessarily bad, it's just that it's limiting. You lose *a lot* of information compared to higher bandwidth interactions with a community that you're governing with.
I agree: small scope/many people can work.
♥ 1
I think it's a function of how much information exchange you need to make a good decision.
♥ 1
Replying to @DinaSrinivasan and @swodinsky, @johnjhorton, @Chronotope
She would totally rock it! Shosh, have you considered a fellowship somewhere? Like, CITP for instance? (Not that I'm biased, but…)
♥ 2
Replying to @tim_libert
I'll believe it when I see it. I've been paying attention and so far I have seen no usage proposed that doesn't create a new chokepoint.