Twitter CEO Jack Dorsey, whose platform has become synonymous with yelling, trolling, and a less-than-stable president who uses it for nuclear annihilation, is in a tough spot: Many if not most of the criticisms of his site are accurate, and Dorsey’s tactic of choice when responding to them is usually sputtering out a bunch of five-dollar words about how he needs to make unspecified changes.
On Tuesday, Dorsey agreed to a Q&A with Recode co-founder and New York Times contributor Kara Swisher. Unfortunately, this all played out in Twitter replies—a format that is ideal for conveying quick thoughts à la carte, but very hard to track in real-time. Just as unfortunately, Dorsey’s interview strategy has not changed. About the only thing he conceded was that he has not been handling his role as CEO well and that major issues plague the entire platform.
Follow below, if you dare:
Oh hai @jack. Let’s set me set the table. First, I am uninterested in beard amulets or weird food Mark Zuckerberg served you (though WTF with both for my personal self). Second, I would appreciate really specific answers. #KaraJack
— Kara Swisher (@karaswisher) February 12, 2019
As @ashleyfeinberg wrote: “press him for a clear, unambiguous example of nearly anything, and Dorsey shuts down.” That is not unfair characterization IMHO. Third, I will thread in questions from audience, but to keep this non chaotic, let’s stay in one reply thread. #KaraJack
— Kara Swisher (@karaswisher) February 12, 2019
To be clear, we’re interested in an experience like this. Nothing built yet. This gives us a sense of what it would be like, and what we’d need to focus on. If there’s something here at all!
— jack (@jack) February 12, 2019
Ready
— jack (@jack) February 12, 2019
Swisher started out by asking Dorsey why the site can’t move faster to address criticisms—which are numerous, but the most high-profile of which are rampant harassment and accusations it, as the Week put it, “intensifies and amplifies pathological social tendencies among those who act within, report on, and write about the political world.”
Dorsey answered by saying that “in the past we were trying to do too much” and is now focused on “prioritising by impact.” He admitted that if he had to grade his own performance, he would give himself a “C.” The reason? Twitter has tried, but it’s “been scattered and not felt enough,” and the site has “put most of the burden on the victims of abuse.”
In other words, even Dorsey admits he’s been performing at a barely-passable level, and that there has been a major failure in keeping the site from descending into invective.
A question we ask ourselves all the time. In the past I think we were trying to do too much. We’re better at prioritizing by impact now. Believe the #1 thing we should focus on is someone’s physical safety first. That one statement leads to a lot of ramifications. #Karajack
— jack (@jack) February 12, 2019
Myself? C. We’ve made progress, but it has been scattered and not felt enough. Changing the experience hasn’t been meaningful enough. And we’ve put most of the burden on the victims of abuse (that’s a huge fail). #Karajack
— jack (@jack) February 12, 2019
Dorsey then rolled out a lot of gobbledygook about how Twitter is trying to find ways to “proactively enforce and promote health”—in other words, shifting the goalposts again into muddled conversations about conversations, rather than the impact Twitter is having right now. Also, Dorsey thinks that if Twitter achieves “proactive” “#health,” then reporting or blocking other users can become a “last resort.”
Putting the burden on victims? Yes. It’s recognizing that we have to be proactive in enforcement and promotion of healthy conversation. This is our first priority in #health. We have to change a lot of the fundamentals of product to fix.
— jack (@jack) February 12, 2019
Agree it matters a lot. And it’s the most important thing we need to address and fix. I’m stating that it’s a fail of ours to put the majority of burden on victims. That’s how the service works today. #Karajack
— jack (@jack) February 12, 2019
First and foremost we’re looking at ways to proactively enforce and promote health. So that reporting/blocking is a last resort. Problem we’re trying to solve is taking that work away. #Karajack
— jack (@jack) February 12, 2019
Second, we’re constantly evolving our policies to address the issues we see today. We’re rooting them in fundamental human rights (UN) and putting physical safety as our top priority. Privacy next. #Karajack
— jack (@jack) February 12, 2019
At this point, Dorsey calmly explained that Twitter sees the most abuse happening in “replies, mentions, search, and trends.” Careful observers might note that this is the whole goddamn site, aside from Moments and direct messages.
Dorsey also wrote that Twitter is “focused on understanding what conversational health means,” which is a way of saying that it does not have any idea how to define its core product (or is stalling on the subject). He also added the platform is trying to figure out “what can we do within the product and policy to lower [the] probability” of Twitter spilling over into offline physical safety concerns. Again, no concrete answers:
This is exactly the balance we have to think deeply about. But in doing so, we have to look at how the product works. And where abuse happens the most: replies, mentions, search, and trends. Those are the shared spaces people take advantage of #karajack
— jack (@jack) February 12, 2019
Good question. This is why we’re focused on understanding what conversational health means. We see a ton of threats to health in digital conversation. We’re focuse first on off-platform ramifications (physical safety). That clarifies priorities of policy and enforcement.
— jack (@jack) February 12, 2019
I mean off platform, offline ramifications. What people do offline with what they see online. Doxxing is a good example which threatens physical safety. So does coordinate harassment campaigns. #Karajack
— jack (@jack) February 12, 2019
Within the service? Likely within replies. That’s why we’ve been more aggressive about proactively downranking behind interstitials, for example.
— jack (@jack) February 12, 2019
No, not a police force. I mean we have to consider first and foremost what online activity does to impact physical safety, as a way to prioritize our efforts. I don’t think companies like ours have admitted or focused on that enough. #Karajack
— jack (@jack) February 12, 2019
Exactly. What can we do within the product and policy to lower probability. Again, don’t think we or others have worked against that enough. #Karajack
— jack (@jack) February 12, 2019
Dorsey also claimed that Twitter takes action against “all we can against our policies,” which is certainly curious, because accounts of the platform’s moderation team turning a blind eye to toxic behaviour that actually is reported is rampant (as are accounts of the system being easily gamed by a technique called brigading, in which trolls mass-report relatively innocuous tweets in the hopes of successfully locking down someone’s account).
He also told Swisher that a reporting-based moderation “doesn’t scale,” which is a convenient excuse for why some of the site’s worst bad actors never seem to get penalised.
We action all we can against our policies. Most of our system today works reactively to someone reporting it. If they don’t report, we don’t see it. Doesn’t scale. Hence the need to focus on proactive. #Karajack
— jack (@jack) February 12, 2019
I see the link, and that’s why we need to put physical safety above all else. That’s what we’re figuring out how to do now. We don’t have all the answers just yet. But that’s the focus. I think it clarifies a lot of the work we need to do. Not all of it of course. #Karajack
— jack (@jack) February 12, 2019
Of course, all this talk about how moderation does not “scale” and Twitter should focus on being “proactive” is quite a convenient way to justify inaction—It’s sort of like looking at a burning skyscraper and saying “Boy, this sure looks too big to handle! What could we have done differently?” You know, instead of putting it out.
Swisher tried to nail Dorsey down on hard examples of what Twitter is doing to end harassment, which ended up being yet another foray into the land of platitudes with a few pit stops at relatively minor tweaks that it has made (introducing a rule against misgendering, downranking some “bad actors,” and the mute function, which was introduced… five years ago).
I think we tried to do too much in the past, and that leads to diluted answers and nothing impactful. There’s a lot we need to address globally. We have to prioritize our resources according to impact. Otherwise we won’t make much progress. #Karajack
— jack (@jack) February 12, 2019
Likely a reason. I’m certain lack of diversity didn’t help with empathy of what people experience on Twitter every day, especially women. #Karajack
— jack (@jack) February 12, 2019
1. We have evolved our polices.
2. We have prioritized proactive enforcement to remove burden from victims
3. We have given more control in product (like mute of accounts without profile pics or associated phone/emails)
4. Much more aggressive on coordinated behavior/gaming— jack (@jack) February 12, 2019
1. Misgendering policy as example.
2. Using ML to downrank bad actors behind interstitials
3. Not too long ago, but most of our work going forward will have to be product features.
4. Not sure the question. We put an entire model in place to minimize gaming of system. #Karajack— jack (@jack) February 12, 2019
Dorsey also responded to accounts of people taking Twitter “breaks and purges,” writing that it “feels terrible” and he doesn’t “feel good about how Twitter tends to incentivise outrage, fast takes, short term thinking, echo chambers, and fragmented conversation and consideration.”
He said that fixing this would require him to “change more fundamentals,” but failed to elaborate.
Feels terrible. I want people to walk away from Twitter feeling like they learned something and feeling empowered to some degree. It depresses me when that’s not the general vibe, and inspires me to figure it out. That’s my desire #karajack
— jack (@jack) February 12, 2019
I also don’t feel good about how Twitter tends to incentivize outrage, fast takes, short term thinking, echo chambers, and fragmented conversation and consideration. Are they fixable? I believe we can do a lot to address. And likely have to change more fundamentals to do so.
— jack (@jack) February 12, 2019
More rambling:
It’s the reality. We tried to do too much at once and were not focused on what matters most. That contributes to slowness. As does our technology stack and how quickly we can ship things. That’s improved a lot recently #karajack
— jack (@jack) February 12, 2019
Dorsey then reiterated that Donald Trump’s infamous account does in fact enjoy above-the-rules status due to his position as president. But he insisted, against all indications to the contrary, that said policy was motivated in any way by metrics:
We hold all accounts to the same terms of service. The most controversial aspect of our TOS is the newsworthy/public interest clause, the “protection” you mention. That doesn’t extend to all public figures by default, but does speak to global leaders and seeing how they think.
— jack (@jack) February 12, 2019
I don’t believe our service or business is dependent on any one account or person. I will say the number of politics conversations has significantly increased because of it, but that’s just one experience on Twitter. There are multiple Twitters, all based on who you follow.
— jack (@jack) February 12, 2019
Dorsey also declined to name a “historic newsworthy figure” that he would ban, which is about the softest ball one could throw:
We have to enforce based on our policy and what people do on our service. And evolve it with the current times. No way I can answer that based on people. Has to be focused on patterns of how people use the technology. #karajack
— jack (@jack) February 12, 2019
Really? Can’t think of anyone?
The thread then just sort of petered out from there, but Dorsey did take the time to say that one of the Twitter accounts he has the most “respect” for is Tesla and SpaceX CEO Elon Musk.
To me personally? I like how @elonmusk uses Twitter. He’s focused on solving existential problems and sharing his thinking openly. I respect that a lot, and all the ups and downs that come with it #karajack
— jack (@jack) February 12, 2019
Musk’s time on the site, of course, has had catastrophic consequences. Those include a weed joke that triggered tens of millions of fines from the Securities and Exchange Commission and his loss of the Tesla chairmanship, as well as a mostly one-sided beef with a cave diver who helped rescue a trapped Thai soccer team—who Musk publicly smeared as a pedophile.
That latter incident resulted in a major defamation lawsuit. In fact, Musk has sent so many erratic tweets that news outlets have publicly discussed his sanity.
This is probably not the example you want to be trotting around! But Dorsey, who seems to have little idea of what he has unleashed but also to be convinced the right amount of word salad can rein it back in, really does not look like he has any idea what he’s doing.
Other than focusing on “conversation,” whatever that means.
Tried to do too much at once. Wasn’t focused on what our one core strength was: conversation. That lead to really diluted strategy and approach. And a ton of reactiveness. #karajack
— jack (@jack) February 12, 2019
This thread was hard. But we got to learn a ton to fix it. Need to make this feel a lot more cohesive and easier to follow. Was extremely challenging. Thank you for trying it with me. Know it wasn’t easy. Will consider different formats! #Karajack
— jack (@jack) February 12, 2019
[Recode]