Is there any concrete evidence that people get better products thanks to A/B testing?
♥ 14↺ 2
December 2022
165 tweets
Replying to @jfhbrook
Any study of actual satisfaction, matching of stated goals for use, etc. Quali studies. Not increased engagement or habituation.
♥ 2↺ 1
Replying to @taubeneck
Well, I got the same joke in both treatments 😁 mstdn.social/@mattmay/10943…
♥ 5
Replying to @StefanBielau and @peterhense
Hence my question.
Replying to @WolfieChristl and @jfhbrook
Well mine for one 😁
But there is an underlying point too: if it doesn't make things better for people, we don't need to support it in the tech stack. This matters as we design what comes next.
♥ 4
Replying to @tim_libert and @avadacatavra
That's why the job specifically requires PowerPoint skills: for the rats that don't fear cats.
♥ 1
Replying to @tim_libert and @avadacatavra
I expect PETA will get involved.
♥ 4
@vinnysgreen Looks like your account has been compromised, got a weird DM?
Replying to @vinnysgreen
Haha sorry, yeah scams all start like that 😁
Torn between wanting England to lose and wanting France to do the honours in the next game.
♥ 4
Replying to @struthious
Same, though I doubt that's enough, if you focus too much on him then there's plenty of room for others who aren't bad either!
♥ 1
I wish I could count the number of times people have told me "you can publish anything you want on the web, that shows it isn't captured!"
The question was never "can you publish" but "who decides whether it gets seen?" *That* is what gatekeeping means.
Quoting a tweet by @micahtredding ↗
♥ 23↺ 8
On the web, we have two variants of a single category of gatekeepers: search and social. They look superficially different - one is pull, the other push - but that's just a UI detail. It's all algorithmic media exposed in slightly different ways.
♥ 8↺ 4
Together, they have full editorial control over what is allowed to matter on the web. That their editorial decisions are automated doesn't make them any less editorial. That they may be personalised is just second-order editorialising.
♥ 6↺ 1
We haven't faced this degree of gatekeeping in centuries, if ever. A tiny group of people decide what's relevant. You don't go to jail for disagreeing with their demands, you're just buried under the boosted signal from those who collaborate.
♥ 11↺ 1
At the end of the day, this is incompatible of course with democracy, but also with human flourishing. If you're in tech, I think it's a good idea to consider if your mind is best used working on some this_is_fine.gif project or on fixing this.
♥ 11↺ 3
Replying to @null and @wjmaggos
The fedi is promising and has been promising for a few years now, but we still have our work cut out to make it capture-resistant. Replacing "everyone can publish" with "everyone can start an instance" doesn't help if someone can figure out gatekeeping anyway.
♥ 1
I'm cautiously optimistic though! I just think that we need to work *now* on locking it open. We shouldn't assume that ActivityPub is the end all & be all.
Replying to @JamesGRobinson
It's terrifying for the moment, but when you look at the authors or books who had to face this it's pretty reassuring in case it happens to you!
♥ 1
Replying to @stdlib
Plenty of good opportunities in this jobs.protocol.ai/companies, including some Rust and other kinds of low-level stuff. Please don't hesitate to DM!
Replacing visible, identified labour with hidden, interchangeable labour in cheaper and less protective jurisdiction isn't a shortcoming of "AI" — it's the whole point. It's a method of globalised domination for which worker movements (and public perception) are unprepared.
Quoting a tweet by @chloexiang ↗
♥ 20↺ 4
This looks like a cool datagov job at @EPFL_en! recruiting.epfl.ch/Vacancies/2681…
♥ 1↺ 1
Everyone is answering about constructor theory but I'm curious to hear if anyone has thoughts on the capability theory side of the question?
Quoting a tweet by @davidad ↗
♥ 3
"Germany’s intelligence services have for years said that the greatest threat to the country came from domestic, far-right extremist groups."
It's the greatest threat everywhere.
Quoting a tweet by @palafo ↗
♥ 20↺ 8
Replying to @StefanoGogioso
I would love to, I think that capabilities might be key to defining good governance and so a formal treatment would be valuable. Alas, I don't think that I have the cat chops (yet) to contribute much, working on it. About to hop on a long flight but happy to chat more if useful!
♥ 1
Wonderful piece from @mariafarrell pointing out that many tech system referred to as "ecosystems" don't deserve that name; they're much more like plantations. The metaphor shift is both precise and much-needed, ecosystems aren't compatible with control.
Quoting a tweet by @mariafarrell ↗
♥ 26↺ 6
Replying to @OlufemiOTaiwo and @NickJBrumfield
Has someone made a list/review of innovative agroforestry? I would love to read that.
♥ 1
Replying to @henryfarrell
Ireland has a DPA?
More seriously, I think that this might fall under ePrivacy, which would mean that this does not have to go through a supervisory authority. The @CNIL has been quite effective at doing that, and I don't think it's even far-fetched here.
♥ 3
Replying to @neilturkewitz and @mariafarrell
That's a fair point but I disagree on two points. I think that we *could* build a tech system with the rich opacity and complexity of an ecosystem. It would probably require new techniques because it's not something we've really tried that I'm aware (not even in ALife).
♥ 1
The thing is that it wouldn't do anything for us, we wouldn't be able to control it to produce anything useful for us. I suspect that there's a no-free-lunch in control vs intelligent complexity. For the same reason if we built AGI it would be hard to get it to do work for us.
♥ 1
The second point is: I believe that a techno-*social* environment could easily have ecosystem complexity! But for that to work, we need to liberate people, to find ways to prevent control from emerging.
♥ 1
Replying to @Elibietti
Lol, I still dismiss the consent dialog that they keep blasting whenever I have to use WhatsApp (which is getting rarer mercifully).
I would argue that it is in sufficient common use to justify its addition on the merits anyway.
Unicode friends: you know what to do.
Quoting a tweet by @Ayegill ↗
♥ 4
Describing armed paramilitary militia showing up to intimidate artists and performers as a "debate" is nothing short of far-right ideological alignment from @nytimes. This isn't both-sidesism; this validates domestic terror.
Quoting a tweet by @BrandyZadrozny ↗
♥ 24↺ 6
Is everyone else being asked to consent to having their DMs read by Bari Weiss?
Something tells me this doesn't quite hold water as valid consent for ePrivacy communication secrecy purposes, no?
♥ 123↺ 31
Replying to @clancynewyork
They've always done some of what this list, but it's weird that this is being surfaced just now. Like, do they even have lawyers left?
For American lawyers, this is "consent" :)
♥ 12↺ 1
Replying to @privacy_spider
Evidently, but they weren't always the plaything of a cabal of remarkably stupid attention-craving ideologues. It makes a difference.
♥ 3
Replying to @clancynewyork
I don't know, it's all speculation evidently, but they've been dropping out of the CSAM fight it seems so I'm not sure that was the trigger. It could really be that Elmo asked counsel about giving access to private messages to his coterie and this is CYA.
♥ 4
Replying to @Chronotope and @clancynewyork
I've had it on two accounts. It's possible that I may be associated with Europe. If Americans don't get it but Europeans do, that's interesting because it might be a really bad attempt at ePD CYA.
♥ 4
Replying to @clancynewyork and @Chronotope
Oh that could totally be the case indeed — it's hard to guess for sure!
♥ 1
Replying to @_matheus_scs and @matthew_d_green
Of course, but I find it interesting that someone in compliance decided that something changed to the point of adding this.
♥ 1
Replying to @RuairiRobinson and @matthew_d_green
Right — which makes it interesting that someone in compliance suddenly decided that this had to be added. What changed?
♥ 1
Replying to @_matheus_scs and @matthew_d_green
For American lawyers, this counts as consent (you don't "have" to click, you could go read a book instead). Don't shoot the messenger! 😅
What's interesting is that people seem to be getting this in the EU but not the US, so maybe they intended this as ePD consent — it isn't.
♥ 2
Replying to @RuairiRobinson and @matthew_d_green
But not the consent notice — someone changed their mind very recently about this.
Replying to @RuairiRobinson and @matthew_d_green
Interesting. This just came up nearly simultaneously in two unrelated accounts, in frequent use. Maybe a microservice died?
Replying to @null and @jbakcoleman
I agree that it's ramping up but merchants of uncertainty has been used in privacy for a while as well as flooding with feedback (DNT). Blaming the victim ("consent") is common across the board. I think they've only had to use this in the few places where pressure showed promise.
♥ 1
Replying to @lilianedwards and @je5perl, @riptari
Right, I was thinking of this bit which I think is what applies?
♥ 2
Replying to @riptari and @lilianedwards
I think it's worse because it doesn't have to be the DPC, it can be a real DPA instead
Replying to @henryfarrell
Ireland has a DPA? More seriously, I think that this might fall under ePrivacy, which would mean that this does not have to go through a supervisory authority. The @CNIL has been quite effective at doing that, and I don't think it's even far-fetched here.♥ 3
♥ 1
Replying to @LuizaJarovsky and @lilianedwards, @charlesarthur, @jasonkitcat, @tomskitomski, @DPCIreland, @riptari
It definitely doesn't look great, in practice this is quite common. I'm not implying that that makes any of it OK, but I can't say that I'm very surprised at any of it.
♥ 1
Spent a most excellent day talking with a group of smart, strategic publishers and interested parties about how data could be governed so that the Internet stops being hostile to privacy and journalism.
It's a tough road ahead but the bright people involved give me hope.
♥ 37↺ 2
Replying to @can
I'll be in CA Dec 22-31, great coordination.
First they ignore you.
Then they laugh at you.
Then they suspend your account.
And then? #joinmastodon
♥ 33↺ 14
I couldn't care less about Space Karen's jet but delighted to go follow it on Mastodon and get a smell of free speech.
Quoting a tweet by @internetrebecca ↗
♥ 2
Having seen how this guy runs a website it's a death wish to ride in his cars, frankly. Never know if that next software update won't kill you.
Quoting a tweet by @AdamParkhomenko ↗
♥ 10↺ 4
Shout out to Manu here who got banned by hurting Elmo's feelings before it e cool.
Quoting a tweet by @trekonomics1 ↗
♥ 6
How long before @Marvel releases an edited version of Iron Man 2 without the scene Elmo paid to be in? It really hasn't aged well.
Replying to @clancynewyork
Like Trump, there's always a tweet. Next, he'll be denying that the bankruptcy is real.
♥ 2
Replying to @MarietjeSchaake
I find it hard to imagine that Post won't have the same problems as Twitter, only faster and possibly more insidious. Aram's analysis sealed it for me.
Quoting a tweet by @Chronotope ↗
♥ 2
Replying to @justinph
I don't know, it does software autoupdates OTA, no? What happens when he went hardcore on some guy and they ship something really bad?
Replying to @dmarti
💯
Folks who are considering Post should take a moment to read this thread. You're just setting yourselves up for the same drama.
Mastodon has rough edges, but it's guaranteed to only get better.
Quoting a tweet by @Chronotope ↗
♥ 23↺ 6
Replying to @null and @john__rosevear, @histoftech
You sound like maybe you didn't read the thread?
Replying to @null and @john__rosevear, @histoftech
Most of the issues listed by @Chronotope aren't technical, and many are specifically about the business model of journalism. Post seems intent on intermediating media, which we've seen over and over. If my livelihood depended on a sustainable press, I wouldn't play along.
♥ 2
Or, you know, if I thought that independence of the press matters. It's understandable that media companies and journalists made mistakes when this was all new. It's distressing to see people just walk into the same trap as if nothing had happened.
♥ 1
I spent five years in media arguing that the privatisation of digital infrastructure was a freedom of the press issue, to be tackled head-on. Too often, people squinted as if I were some kind of weird activist.
Reporters are now being censored on a major channel. #twitterpurge
♥ 58↺ 15
I'm not saying this to pat myself on the back; I likely didn't make the case well enough. But hoping that things will magically work out if media businesses just stay neutral isn't a strategy.
♥ 8↺ 1
And hoping that the exact same setup in ads, search, browsers, OSs… will somehow work out better is naive. The question is when, not if.
♥ 7↺ 1
Replying to @null and @MadronaCitizen
Preaching to the choir my friend :) But it can lack the air of propriety required to carry weight in some circles.
As you may recall:
• MIT partly hosts W3C, but withdraws Dec 31
• A W3C nonprofit needs to take over on Jan 1
• MIT has assets (dues, contracts, IP…) to transfer to new W3C
• I was elected to the W3C Board & in the negotiations
Two weeks from cutover it's going… poorly. 🧵
♥ 519↺ 213
On November 17 we saw terms that were unfair in that they made W3C carry the cost of MIT’s questionable accounting practices and failed to honour prior agreements, but that came with sufficient funds to support the continuation of W3C Inc.
♥ 54↺ 3
Right before the Board's December 10-12 meeting, we received a draft contract from MIT with a worse offer.
It transferred liabilities much higher than the assets that came with and significantly decreased the funds that W3C would receive.
♥ 50↺ 1
Replying to @lolaodelola
That pretty much sums it up, yeah.
Replying to @neilturkewitz and @knightcolumbia, @heidikitrosser, @jamalgreene
I hadn't, thank you! Throwing them on the queue.
♥ 1
Replying to @lolaodelola
Thanks! In terms of stress, it's the MIT/W3C team I'm worried about. Me, I'm not stressed, I'm just fucking mad.
♥ 1
Replying to @johngineer
Well, this is an aspect that I'm not an expert on, I don't know how much having a host in the US grants some influence. But I think that they would be worried with having US companies depend directly on eg. a Chinese host for governance.
♥ 19
Replying to @crislar and @MIT
I think it's been established that reputation isn't something that MIT cares a lot about…
♥ 28↺ 1
Replying to @svgeesus
I can't discuss that here, but a key problem is time: how do we hire the MIT staff in the meantime and how do we guarantee that the systems MIT owns continue to operate?
♥ 4
Replying to @lmrlaurent and @EDRLab_org
Thank you! Yes, I'm pretty angry too.
Replying to @ventilacka
I completely agree with your observation, but this isn't my area of expertise. I know a few books on standards, but they don't cover that area, sorry.
♥ 1
Replying to @photomatt and @WordPress
Thanks a lot for the offer, much appreciated. Right now the issue is financial and timing. I hope that we can resolve it without corporate sponsorship but that might be optimistic…
You're not members right? After this storm, I'd love to chat about that bit if you're up for it.
♥ 16
To be clear, I'm not selling anything. It's just that Wordpress is a pretty damn big thing for the web, and I'd like to understand if you think governance should work differently, how, etc.
♥ 9
Mbappé Mbappé Mbappé
MBAPPÉ MBAPPÉ
MBAPPÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉ
♥ 12↺ 1
You can trust the French to show up 80 minutes fashionably late to a game. #ArgentinaVsFrance
♥ 9
Replying to @__apf__
Depends on the owner and conditions, but hell yeah, I think it's one of the few big tech companies that has some potential future plan which I find attractive. We could really build something good here.
♥ 6
Are there articles or books you'd recommend on #biodiversity? I'm mostly looking at foundational/conceptual, the basics I'd need to ask cogent questions from people who work in conservation, ecology, etc. I'm working "something like" biodiversity and would like to get it right.
♥ 3↺ 1
If you are on both, I'd appreciate answers on the fluffy site: https://mastodon dot social/@robin/109538423109645874
♥ 3
Replying to @theseniorbarn and @s__ballesteros
Thanks!
Replying to @jlivingood
I won't comment for the time being but on general terms this is constructive and I appreciate it!
♥ 1
Replying to @danbri and @PryvitKyle, @OR13b, @w3c, @dckc, @timberners_lee
Basically what Kyle said. I think there's useful stuff in DID space but the dust needs to settle some. I don't know if IPID is still active, I mean to find out.
There's IPFS stuff in Chromium and Kyle can talk about the Brave bits I'm sure :)
♥ 3
Replying to @PryvitKyle and @danbri, @OR13b, @w3c, @dckc, @timberners_lee
Yeah, we're due to revisit bundling, apps, intents I reckon. I think that we can make better mistakes than the previous times. Stuff like IPFS and UCANs definitely add interesting properties that the Web could use, IMHO, and we can ignore the "sea of shit" as you put it :)
♥ 2
Everything about Post screams that it's the worst of Silicon Valley wrapped up in a single app. Don't make it matter.
Quoting a tweet by @TaylorLorenz ↗
♥ 13↺ 2
Replying to @PryvitKyle and @danbri, @OR13b, @w3c, @dckc, @timberners_lee
I can reach the author, thanks — I'll just get around to it after [*gestures around at the whole thing*].
I'm really looking forward to what Brave does there.
♥ 2
Replying to @PryvitKyle and @danbri, @OR13b, @w3c, @dckc, @timberners_lee
😁
At the same time, reviving dead specs is a great hobby! It should be a thing necrostandardser.
♥ 1
Replying to @PryvitKyle and @dckc, @danbri, @OR13b, @w3c, @timberners_lee
Right, it needs enough overlap to be, well, a protocol. I think if you squint maybe it's there?
I'm having #GroundhogDay conversations in other parts of the stack (mime sniffing is one), maybe we should have a group.
♥ 1
I believe it's @swodinsky who told me that tech execs should get the same scrutiny as politicians. Choosing how to change the world is politics and needs to be treated as such.
Too much of tech coverage is just vaguely contextualised press releases.
Quoting a tweet by @libshipwreck ↗
♥ 56↺ 15
Replying to @OR13b and @danbri, @dckc, @PryvitKyle, @w3c, @timberners_lee
It cuts both ways, too. Things that should be Recs can get endlessly stalled. It becomes a bit like a bank where you can only get a loan if you can prove you don't need it.
♥ 2
I'm delighted to report that MIT and the W3C have reached an agreement and that the W3C will kick off 2023 with funds and assets.
Many thanks for all your support and to all those who worked towards reaching an agreement. Enjoy the holidays!
♥ 182↺ 52
Replying to @mishmosh
Oh yes, lots of that 😁
Replying to @garrett_wollman
You should see directly with the W3C team for details!
🤔
♥ 10↺ 1
Does this opinion piece advocate a coherent position grounded in fact? No, but is it arguing a novel position that hasn't been presented dozens of times before? Also no.
The Signal App and the Danger of Privacy at All Costs nyti.ms/3vnNHTc
♥ 17↺ 2
Replying to @redteamwrangler and @evacide
OMG this is so funny.
Replying to @bertil_hatt
Yeah. I have no idea why anyone would consider this relevant to debates around privacy.
♥ 1
I have a 13yo and a 9yo, both smarter than their dad, and I'd like to help them develop the moral skills they need to live the good life. Are there any good books about virtue ethics and habits for roughly that age?
♥ 21↺ 3
Replying to @codefreechris
I understand that, but having the conceptual apparatus on top of experience helps reflect on what to do and why. I can't be all their phronimoi.
Also, I think you have it backwards: being virtuous is foundational to being smart.
Replying to @wheresaddie
Of course, I'm not trying to delegate to a book! However, I think that having the intellectual tools *as well* can really help, especially when harder decisions come up or when you need to understand how to change.
Replying to @samuelgoto
I agree, but I'm looking for something they can pick up that isn't the Nicomachean 😁 I've also quit the Kindle and it's true that it certainly helps with the curiosity part!
♥ 2
Replying to @codefreechris
The fact that they're already smarter limits my ability to serve as model since that means that they already exceed my ethical competence and practical wisdom in many ways. Understanding this relationship is precisely why the conceptual apparatus matters.
Replying to @codefreechris
As I've said already, I don't think that you're operating from a good definition of smart. But you don't seem interested in doing much more than repeating yourself, so have a great day!
Replying to @Elibietti
Yes, I loved that when I was a teenager! It's not specifically what I'm looking for in this case, but that one is clearly one I'd like to make sure they get to.
♥ 2
Replying to @1upm and @Chronotope
I'm not looking for high brow! Thanks, this is intriguing, I'll check!
♥ 2
Replying to @mberk
Thank you! I've seen Alasdair MacIntyre referenced many times but haven't read, I'll check!
♥ 1
Replying to @samuelgoto and @sivers
I'm not but it does look compelling, thanks! And Wonder worked great for both, much agreed!
♥ 1
Replying to @sivers
It does look pretty good indeed, thanks!
I tried Signal Stories and you know what? They're great. Building that kind of feature on top of open protocols is what we should be doing a lot more of.
♥ 26↺ 4
Replying to @sanjuktampaul
Like, Stories but you actually trust the product. It's a weird feeling!
♥ 2
2022 was a long and demanding year, but one which leaves me hopeful and ready.
♥ 15
One reason to celebrate is that this year I got to meet many people who I'd been speaking with on the Internets for the longest time but had never met in the flesh.
It had been too long and I cherish those encounters! ♥️
♥ 3
It was a year of loss. First, we lost our beloved cat Mochi.
♥ 4
Right - why would anyone want to make security attractive to laypeople?
Quoting a tweet by @cyberflamingo ↗
♥ 2