\n
LY:<\/strong> Particular to your comments about people being force-fed certain types of media that have certain types of bias, I think we often underestimate people\u2019s ability to seek out the real information and actually make the decisions for themselves.<\/p>\n<\/div>\n\n
LY: <\/strong>And that\u2019s the beauty of how this works, right? The power of the platform, right? To be able to seek out that information. That\u2019s when Community Notes comes in to give you a little fact-check from people on the ground. And as we\u2019ve seen, particularly in the last 20 days, how Community Notes has become so powerful, so meaningful, and now that you have so many new readers and contributors through Community Notes, and it\u2019s quite a unique capability that X offers.<\/p>\n<\/div>\n\n
EM:<\/strong> Community Notes started off as Birdwatch. It existed before the acquisition, but it was in a very nascent form and just limited to, really, the US. I put a massive emphasis on this. This is something that we need to make a top priority at the company.<\/p>\n<\/div>\n\n
LY:<\/strong> Forty-four countries and over 100,000 contributors. Growth has been really just explosive. So it\u2019s great.<\/p>\n<\/div>\n\n
EM: <\/strong>I\u2019m always amazed at the sort of batting average of Community Notes. Nothing\u2019s perfect, but if a community note has been voted in and stays, the batting average is high 90s with accuracy. In fact, it\u2019s the most accurate thing. If I want to know what\u2019s going on, I\u2019m going to look at the community note to figure out if it\u2019s real or not.<\/p>\n<\/div>\n\n
LY:<\/strong> I think it\u2019s really important to bring up because it\u2019s one of the singular, most powerful pieces of feedback that we get, even from our biggest skeptics, is our approach to what we\u2019re doing here that is completely transparent for all to see. When you\u2019re talking about Community Notes, I think, in the last 20 days, I\u2019ve probably posted five or six times about the evolution and the sophistication and maturation of Community Notes alone. But the way this company is innovating with complete transparency may invite some more of those skeptics. But like no other company in real time, tell them exactly what\u2019s going on. And it\u2019s been really powerful, positive feedback.<\/p>\n<\/div>\n\n
\n
\u201cCommunity Notes will be the single best source of truth on the internet, period.\u201d<\/p>\n<\/div>\n<\/div>\n
\n
EM: <\/strong>Community Notes is infinitely better than some sort of mysterious censorship bureau where nobody knows who they are or why they make decisions or anything. That\u2019s the kind of thing that lends itself to manipulation of information. I want to be clear, Community Notes will be the single best source of truth on the internet, period.\u00a0<\/p>\n<\/div>\n\n
I want to just go through the list quickly just to acknowledge the great work that has been done. We have picture in picture for videos on web. A ton of video features. We\u2019re rapidly reaching parity with YouTube and may exceed them. I test it with an occasional video game that\u2019s livestreamed, and we\u2019re also synchronizing the live video and the stored video because they really should be the same thing. A video that\u2019s being uploaded or a video that\u2019s live with a delay are essentially the same thing. So we\u2019re making sure the interface is going to be the same.\u00a0<\/p>\n<\/div>\n
\n
I also want to acknowledge those who have deleted code. My general philosophy on software is that you get one point for adding a line of code and two points for deleting. This is a really big deal. So thank you to everyone who has worked hard to simplify the code base and delete and simplify the software.\u00a0<\/p>\n<\/div>\n
\n
We are working toward money transmission licenses. Hopefully, those will come through in the next few months. We\u2019ve added ID verification, the ability to hide likes. We made it clear that we will support someone if their employer picks on them or fires them for what they post on the system. We will provide them with legal support.\u00a0<\/p>\n<\/div>\n
\n
We\u2019ve encrypted DMs, and that\u2019s also something that we will expand upon. The acid test for DMs, as with Signal, is that if somebody has a gun to my head, I still can\u2019t tell you what somebody\u2019s DMs are. Basically, they need to be definitively private or you really can\u2019t trust them. The team has done a lot to reduce child exploitation material. That has gone down dramatically, I think somewhere on the order of 90 percent-plus.<\/p>\n<\/div>\n
\n
LY: <\/strong>Not to interrupt you, Elon, certainly. But it is one of the biggest areas of support we get from the partner groups or the interest groups that we work with specifically in this area of what it was like a year ago on the platform and our philosophy of zero tolerance on the platform now. So it\u2019s something really important that we should celebrate.<\/p>\n<\/div>\n\n
EM: <\/strong>Totally. There are a ton more. We\u2019re also working to improve search dramatically. We want to have a semantic search. So if you type in search terms, it knows what you mean and then can bring up the text, pictures, and video that are what you mean. Not just strictly a text comparison. We\u2019re seeing the beginnings of that with the \u201csee similar posts\u201d capability. That is sort of an AI-based system.\u00a0<\/p>\n<\/div>\n\n
Really, what we need to move toward is the entire system, all the \u201crecommended for you\u201d posts being AI-based. Essentially, you populate a vector space around the user and then any given post would also have a vector space. You correlate the two vector spaces, and you show people what is most likely interesting to them. And that will update all of the heuristics that we currently have.<\/p>\n<\/div>\n
\n
We went through a similar process at Tesla with Full Self-Driving where it started off with almost entirely C++ code and then gradually became more and more a neural net. And now, the latest version of Full Self-Driving, called Rodeo, is entirely neural nets. I think that\u2019s going to be very compelling for the recommended feed. It\u2019s going to be a game-changer. And it will also mean that if somebody posts a very interesting reply, normally that is buried by default, that is buried with very little exposure, but sometimes the reply is more interesting than the original post. And we currently lose all that. With a purely AI-based system, whatever the reply is, if the reply is compelling, that will be shown more prominently than the original post if it\u2019s compelling.<\/p>\n<\/div>\n
\n
LY:<\/strong> Elon, can we talk a little bit about potential continuing efforts that we really are leading the marketplace in, in fighting spam and bots? Because when you look at our user and performance numbers, those are net of everything that we\u2019ve done aggressively fighting the spam and bots.<\/p>\n<\/div>\n\n
EM:<\/strong> Yeah. Where things are today is that even generally available AI, very basic AI much less the sort of cutting-edge stuff, is easily able to overcome the various \u201care you a human?\u201d tests. Like, solve this puzzle or tell me which of these pictures have a traffic light. AI is faster and better than people at identifying and passing these \u201care you a human?\u201d tests. You don\u2019t even need to initially have bots.\u00a0<\/p>\n<\/div>\n\n
You can have, you know, 100 people in a warehouse, each with 100 phones, that are all just going through the \u201care you a human?\u201d tests because they are human, and they actually will also be way better than the average human because they do it all day. This is a really big deal, addressing the bots and trolls. It\u2019s important to avoid platform manipulation, where someone will spool up 100,000 bots or 100,000 fake accounts to make something seem more viral than it really is or to just spam users with crypto scams, or whatever the case may be. It essentially adds noise to the system and makes our platform less useful.\u00a0<\/p>\n<\/div>\n
\n
\n
\u201cAdvertisers should not find bots very appealing unless the bots are actually buying, which they won\u2019t\u201d<\/p>\n<\/div>\n<\/div>\n
\n
It\u2019s certainly possible if you turn a blind eye to bots for something to seem far busier than it really is. And so, there are other things that may seem incredibly busy, but they\u2019re just really bots. But obviously, bots don\u2019t buy things, so advertisers should not find bots very appealing unless the bots are actually buying, which they won\u2019t. And then a bunch of the bots will be spam. That means there is bad content. Content that is not interesting. I regard bots and trolls as reducing the signal to noise of the collective mind that is the X platform.<\/p>\n<\/div>\n
\n
LY: <\/strong>And safe to assume that the Not a Bot program will expand at the right time beyond Singapore and New Zealand, right?<\/p>\n<\/div>\n\n
EM:<\/strong> Yeah. We\u2019re running an experiment to see what the consequences would be if there\u2019s literally $1 a year, which is 0.3 cents per day. Sometimes, I get this absurd thing where it\u2019s like, \u201cHow will people in poor parts of the world afford?\u201d I\u2019m like, \u201cThey are somehow on the internet. They have an electronic device, and they are on the internet, and a dollar a year is 0.3 cents a day.\u201d This is affordable for anyone, obviously.\u00a0<\/p>\n<\/div>\n\n
My experience from X.com way back in the day \u2014 X \/ PayPal \u2014 was that the spammers and fraudsters are the ones who complain the loudest. Whenever some action was taken, it would get the most righteous indignation from the fraudsters because they don\u2019t like being shut down. So we expect the same thing here again. But we are being careful about this. That\u2019s why we didn\u2019t just start worldwide. Just trying it out in the Philippines and New Zealand. So, we\u2019ll see how it goes [and] make some adjustments.\u00a0<\/p>\n<\/div>\n
\n
Even $1 a year, while obviously a tiny amount of money for anyone, any human, is still expensive if you want to do 100,000 bots. Now, you have to have $100,000 to get 100,000 bots, and you need 100,000 payment methods, which is even harder. We have to proceed carefully on that front.\u00a0<\/p>\n<\/div>\n
\n
LY: <\/strong>And when we think about the hiring business ready to be disrupted with X Hiring. Also when you think about the PR business with X Wires, [it\u2019s] pretty exciting, too. Approaching something that I don\u2019t think many people ever have thought needs, you know, contemporizing, right?<\/p>\n<\/div>\n\n
EM: <\/strong>Yeah. Really, there\u2019s no need for PR Newswire. We are PR Newswire. How many individuals actually subscribe to PR Newswire? Basically no one. If you actually want to put out information as a company, we\u2019re by far the best method of doing so. That\u2019s obviously something we just need to highlight to companies. Corporations that have been around for a while do the same thing they\u2019ve been doing for a long time until we point out, \u201cHey, there\u2019s a lower cost, easier way to do it, which is just put it on the platform.\u201d And that\u2019s what we\u2019re going to do.<\/p>\n<\/div>\n\n
LY:<\/strong> I should get the numbers and the result because we just had two pretty nice size agencies merge: Camelot and PMG. They came to me the night before. I don\u2019t know if they had heard this weeks ago. They asked us to help them launch the announcement of the merging of the company. They experienced a response that was, you know, just 10x what they could ever imagine going from journalist to journalist or publication to publication to spread the word out. So I think we\u2019re onto something.<\/p>\n<\/div>\n\n
EM: <\/strong>Sounds good. Cool. Yeah, forgive me if I\u2019m forgetting to mention because there are so many things that have been accomplished in a year. I think this is the fastest rate of innovation maybe ever for any internet company. Maybe X \/ PayPal back in the day was similar, but other than that, it\u2019s certainly a faster rate of innovation for any mature company that has been around for a while by far.<\/p>\n<\/div>\n\n
LY: <\/strong>I think the rolling list of numbers is somewhere around 200 product innovations. I\u2019m not sure in most companies\u2019 lifetimes or life cycles that they can talk about those. And even as, you know, consequential as the rollouts have been here, it\u2019s just one year. And I think that\u2019s why \u2014 because we\u2019re pushing so much against legacy \u2014 people are paying a lot of attention to what we\u2019re doing.<\/p>\n<\/div>\n\n
EM: <\/strong>Indeed. The time to be concerned is not when, say, there are media stories about our company dying but when they don\u2019t bother to write the stories.<\/p>\n<\/div>\n\n
LY:<\/strong> Right. If they stopped writing about us, then we have become irrelevant. Or we always say we\u2019ve become quite boring, and I don\u2019t think we\u2019re boring. [Laughs]<\/em><\/p>\n<\/div>\n\n
\n
\u201cThe time to be concerned is not when, say, there are media stories about our company dying but when they don\u2019t bother to write the stories.\u201d<\/p>\n<\/div>\n<\/div>\n
\n
EM: <\/strong>There\u2019s also longform content. We can do longform posts now with actually pretty good editing in the posts. We\u2019ll be releasing the ability to have a very sophisticated text output comparable to what someone would do on some of the other platforms out there. Basically, complete rich text posting that [will let you] post a novel, basically, including an illustrated novel with video. Anyways, there\u2019s a ton. We are leaving out a few things. That\u2019s a partial list of accomplishments. There\u2019s a lot more to come in the future, obviously.<\/p>\n<\/div>\n\n
LY: <\/strong>We\u2019re all looking forward to that. There is no way any of this long, long list of accomplishments could have happened without the unconditional dedication of this team. It\u2019s one of the biggest things I\u2019ve experienced and am in awe of since I joined the company.\u00a0<\/p>\n<\/div>\n\n
A company that was 8,000 [employees] has gone to a little under 1,500 and accomplished what we have accomplished. I think it\u2019s pretty much writing a rule or a playbook that other companies should pay attention to. So thank you to everyone. It\u2019s incredible output.<\/p>\n<\/div>\n
\n
EM: <\/strong>Super exciting. I look forward to doing the next one live on our platform.<\/p>\n<\/div>\n\n
LY:<\/strong> That\u2019ll be next quarter. And we can talk about a lot more then. Alrighty.<\/p>\n<\/div>\n\n
EM: <\/strong>Are there any questions? I\u2019m happy to just keep going.<\/p>\n<\/div>\n\n
LY:<\/strong> Yeah, if we send them to Elon, we\u2019ll answer them.<\/p>\n<\/div>\n\n
EM:<\/strong> Just reply to my post<\/a>.<\/p>\n<\/div>\n\n
LY: <\/strong>There was one question from someone I thought was interesting of what they haven\u2019t heard about yet: what in your mind could come next that\u2019ll surprise everybody?<\/p>\n<\/div>\n\n
EM: <\/strong>I think payments and just generally being a financial hub is going to surprise people [with] just how powerful it is. I really understand this world, obviously. The X \/ PayPal product roadmap was written by myself and David Sacks, actually, in July of 2000. For some reason, once it became eBay, not only did they not implement the rest of the list but they actually rolled back a bunch of key features, which is crazy. PayPal is a less complete product than what we came up with in July of 2000, so 23 years ago.\u00a0<\/p>\n<\/div>\n\n
\n
\u201cWhen I say payments, I actually mean someone\u2019s entire financial life\u201d<\/p>\n<\/div>\n<\/div>\n
\n
When I say payments, I actually mean someone\u2019s entire financial life. If it involves money, it\u2019ll be on our platform \u2014 money or securities or whatever. It\u2019s not just, you know, send 20 bucks to my friend. I\u2019m talking about, like, you won\u2019t need a bank account.\u00a0<\/p>\n<\/div>\n
\n
LY: <\/strong>And that we see coming full opportunity in \u201924, right?<\/p>\n<\/div>\n\n
EM: <\/strong>Oh, yeah. I mean, it would blow my mind if we don\u2019t have that rolled out by the end of next year.<\/p>\n<\/div>\n\n
LY:<\/strong> Someone asked me, which was interesting. I\u2019ll take a shot at it first, and maybe you riff on it a little bit. And it was, \u201cWhat is the one thing that we should stop doing in year two and one thing we should start doing in year two?\u201d And I thought about it. Everything we\u2019ve talked about today. But for people who don\u2019t see our vision yet, people who don\u2019t see what\u2019s happening here at X, stop giving any of that oxygen. Don\u2019t pay attention to it. Heads down. See what\u2019s happening here and realize everything that we\u2019re building here. So we make sure payments makes it, you know, actually as robust [an] offering as possible.\u00a0<\/p>\n<\/div>\n\n
Because you have to focus on what we\u2019re trying to realize: the full vision of X. You gotta lean in and own it. Keep the noise out and just pay attention to what we\u2019re building here. I think it\u2019s good advice for all of us because what we\u2019re doing is something pretty special here.<\/p>\n<\/div>\n
\n
EM: <\/strong>Yeah, absolutely. Just heads down. Focus on improving the product.\u00a0<\/p>\n<\/div>\n\n
One thing I was going to mention. This rule is sort of implicitly implemented at SpaceX and Tesla, but I\u2019m going to be much more explicit about it and also apply at our company, X Corp. In any given meeting, make sure there\u2019s at least one piece of bad news. You can have more than one piece of bad news. If you\u2019re in a meeting with me, try to bring up at least one bit of bad news or more than that. My meetings at SpaceX are really good in this regard. They lead with the bad news. Generally, bad news should be loud, repeated, and immediate. Good news can be said quietly once. So just consider it mandatory to at least have some bit of bad news in any given meeting, and start with that. Another way to think of it is we\u2019re looking for a feedback loop that is as good as possible, and sometimes it is uncomfortable to bring up bad news. We need to make it explicit that bad news should be brought up in every meeting, not just with me or Linda, but in general within the company. Basically, what do we need to improve and make better?<\/p>\n<\/div>\n
\n
\n
\u201cWe need to make it explicit that bad news should be brought up in every meeting\u201d<\/p>\n<\/div>\n<\/div>\n
\n