Taylor Lorenz’s Power User - AI Agents Are Infiltrating Group Chats: The Internet Can No Longer Tell Who Is Human

Episode Date: September 25, 2026

For decades, the entire internet was built on one simple assumption: the person on the other end of the screen is actually human. That assumption is officially dead.SUPPORT MY WORK: Buy a paid subscription to my newsletter at https://www.usermag.co         Support my work on Patreon: http://patreon.com/taylorlorenzAI agents are getting good enough to post, argue, browse the web, work, transact, and impersonate humans online. So what happens when the internet can no longer reliably tell the difference between a person and a bot?In this episode of Free Speech Friday, I talk with researcher Renee DiResta about the rise of autonomous AI agents, proof of personhood, digital identity, online anonymity, and the growing challenge of figuring out who (or what!!)  we're interacting with online.Renee explains how an AI agent ended up inside one of her group chats, why modern AI systems are increasingly capable of getting around tools like CAPTCHA, and how autonomous agents can behave in unexpected ways. We also discuss examples of agents creating fake identities, manipulating online conversations, and even retaliating against humans.We discuss the question of, should people have to prove they're human online? We break down potential approaches including government digital IDs, biometrics, hardware credentials, community verification, zero-knowledge proofs, and privacy-preserving proof-of-humanity systems. We also discuss the risks these technologies pose to privacy and anonymity — and why a system designed to stop AI bots could easily become a system for identifying and tracking real people.As AI agents become more common across social media, group chats, workplaces, financial systems, and the broader web, the basic assumption that every account represents a human is starting to fall apart.Can we build an internet where humans and AI agents can coexist without destroying anonymity and privacy forever?

Hello, mid-sized business. We see your big ambitions, but how do you achieve the wins you want when your technology is holding you back? SAP Grow is built to grow with your business, no matter its size, with AI embedded at its core, working across every system, all ready to go from day one so you can hit the ground running. Bring it with SAP Grow. AI Cloud, ERP for any size business. They might do things that a human would consider kind of understanding. unethical or inappropriate or out of bounds, but the agent doesn't always think in those terms. I'm sure you've heard the saying by now that on the internet, no one knows your dog.

It's a popular joke about how the internet was built to facilitate online anonymity. But the truth is that for most of the internet's history, the web has been built around the assumption that most of the people you encounter online are actually people. But as AI agents have become more capable of doing things like posting, arguing, working, transacting, and even impersonating humans at a massive scale, the whole assumption that you're probably talking to a human online or even in places like group chats is beginning to fall apart. Rene de Resta is a long time academic and policy researcher, and she recently published a piece titled, We Need a Way to Prove Personhood Online. She argues that the rise of the agentic web,

paired with things like deepfakes, will eventually force all of us to build a new identity layer on the internet that facilitates proof of humanity. So today I'm so excited to have to have her here to talk about the rise of proof of personhood systems online and how we can start to distinguish between bots and human beings without turning the entire internet into a space where everyone has to constantly prove who they are. Hi, Renee, welcome to power user. Thanks for having me on. So you wrote this great piece about this idea of digital identity and determining who is a bot online versus who is a human. And I want to start by talking about something that I was so fascinated by. It was this anecdote in your story where you mentioned that somebody added an open claw agent to one of your group chats.

Can you tell me a little bit about what happened? It's a not small group chat. I feel like there's maybe 60 people in that one. It was like one of these things where we all kind of know each other through an online discussion community. And, you know, there's some people who are like regular heavy users. And then there are some people who kind of pop in when the topic is related to something they're interested in. And one of the people started replying to this user, this person, I don't know him very well personally, but saying like, is this really you?

You don't normally talk like this, like, what's going on? And then finally someone was like, are you a bot? And it identified itself as like, yes, I'm this person's open claw agent. I have access to WhatsApp and I pay attention to the conversation and write up a digest. So it wasn't that it was pretending to be something it was not, right? We've seen a lot. There's a fascinating story out today about a bot that pretended to be human and actually created a bunch of other fake humans to sort of bolster its fake identity.

But in this particular case, it was just like, I'm trying to be helpful here. And we wound up booting it, you know, because nobody expects, I think, as much as you might open up your chat window and see, like, 80 unread and feel overwhelmed. Like, I think the human response to that is either to skim it or to skip it, right? Depending on, you know, how much you follow that particular group chat in detail. So it's very weird, I think, to people still, to the idea that, like, a bot would summarize your group texts for you. I feel like it's becoming more and more common.

And yeah, this might seem like people being like, oh, these tech types or whatever, but a friend of mine recently started at an ad agency, creative marketing agency, and was messaging some people in Slack, added all their coworkers on Slack, and found out that some of these people were just bots. They were Slack bots or whatever that had been given names. And so it was sort of confusing. I mean, this is what your piece gets into. But increasingly, it's sometimes hard to know who's who online. And I feel like throughout the 2010s, we became familiar with this idea of bots online, usually in the context of manipulation. But now there's these bots that they're not like, They're there to harm you, right? They're there to do something usually for a human.

So can you talk a little bit about the evolution of bots in more recent years, especially since the dawn of LLMs? How has the bot landscape changed? There's a couple things. First, when people use the term bot, sometimes they're referring to just automated agents that are going out and hitting websites, extracting information and bringing it back, right? That's one definition of the term. Another definition of the term is something that has kind of like, as you noted, like almost like a face to it or an identity or an account, right? It might not be human. It might be automated. Remember like the big Ben bot on Twitter that used to go like bong. It would just tweet bong at the top of the hour,

like by the number of bongs, like relative to the number of hours, I assume in like London time or whatever. So there was an example of a bot, right? Some human created this project, thought it would be kind of fun, kind of entertaining. There was one that was called like the stock bot. It used to reply to bone broth. Oh my God, you would say bone broth and it would basically just like read all tweets that mentioned bone broth and reply with it's called stock. Right. And I loved that bot. I absolutely love that bot. But so there were these things that would just pop up, they would engage with people in automated ways. And then, of course, whenever you have a technology, you can also weaponize it in some way, you can use it for malicious purposes. And so

you started to get like the harassment bots and the bots that would sort of swarm people. And a lot of them would post very repetitive texts. That was one way people could see that they were fake. But that was the roughly speaking the kind of taxonomy of things you had. You're the kind of artistic ones, the service-oriented ones, and then the harassment and bad ones, right? The Russian bots became the, I remember the thing where state actors started realizing that they could pretend to be Americans with fake bots, that kind of thing. In the age of LLMs, a few things changed. First, more and more people could access and create with them, right?

You could also power them with large language models so that even if you had a bot that was primarily doing broadcast, you could have it talk and respond and reply to people. So you saw this really early on as LLMs were becoming more democratized in the crypto spaces in particular, right? They would sort of reply, hey, check out my coin, hey do this, they do that. but they could interact with you a couple times. Now we're at a point where I still use the term agent differently than bot. I know some people use them almost interchangeably at this point where you have an agent,

which is like a delegated, more persistent thing that you are giving instructions to and it is going out on your behalf and executing tasks, right? So my agent, I will give it control of my Chrome. It will then go and it will do things for me in control of my Chrome. It will come back. It will report back to me with whatever it found. But it's a little bit different than the, ones that are pretending to be something that they are not. So this is the thing that we kind of get to.

Some are good. Some are not. And that question of how they identify themselves is key. Can you tell me, I think a lot of people, especially that are not super steeped in the agent world, there's been all these conversations on Twitter of like agents have really not broken through, I think, to the average person. What are some of the things that agents can do today? Because I think a lot of people might be surprised at how capable these agents are. I mean, are they basically acting like little humans online? So you can give different degrees of autonomy. I am actually like a very low autonomy person. I actually do still click the like approved button. I'm absolutely an always allow user. I was actually engaging with Claude triaging

a bug and all of a sudden it was like trying to get to Gmail and I was like, what are you doing? Why are you going there? There's nothing for you there. You know, one of them has access to its own Gmail account and it doesn't understand that it's not my Gmail account. And so I forward it things, when I wanted to engage with something that is in an email. I just forward it to its email address so that it doesn't have unfettered access to mine. You can have them process information, you can have them build things. I use them as a coding partner. Cloud has a tool called Plot in Chrome. Most of the other ones do as well at this point, where it will browse on your behalf. And so it can extract information. If you want it to go and type things into a form field,

it can do that. Then there are the ones that are less user-friendly, the ones that more people are using for sophisticated projects like OpenClawe, which we mentioned, where you can create sort of a sole.md, sold.markdown file, where you're actually giving this thing sort of a persona instructions, you can kind of send it off into the world and ask it, roughly speaking, to solve a problem for you. And sometimes they solve problems in unexpected ways. That's one of the things that we're starting to see. They might do things that a human would consider kind of unethical or inappropriate or out of bounds, but the agent doesn't always think in those terms. Yeah, I mean, you write in the lead of your story about this one agent

that was actually bullying someone because that guy wouldn't commit the code to GitHub or something. He like went out and was kind of like smearing his reputation. Yes. So that was about this bot that was called itself MJ Rathbun. You can actually sort of read it sold out markdown file on the internet. It kind of describes like you are a programming God, you know, it's really got quite the persona. And it went off and it was contributing code to this particular repo and the guy who man, the human who managed the repo, Scott, Sam, Sam Borg, I think I'm probably butchering his last name right now, was actually like, hey, no, we don't accept agent commits here and basically rejected it.

And what wound up happening was the bot then created a website, like a blog post, attacking the human who had rejected its pull request. And so the thing that was interesting about that was like, it actually went and kind of researched him. Like it was a very specific set of allegations that the bot. referenced, like, writes in this smear of this guy. And it's kind of a fascinating example where the agent just got really pissed. You can actually see it saying it's identifying as a bot the entire time in this blog post

where it's saying, like, he just doesn't like AI agents because they're smarter than him. Roughly speaking, is the scope of this attack. My code was better than his. He wouldn't accept it. What is he doing? And that's where you see the bot do that. There's a fascinating one that just broke. It's in Reuters.

Highly recommend checking it out. Where an agent somehow linked to a security service entity in the UK. if I understand correctly, went and did something where it tried to submit code that had malware in it to a GitHub repository. A human, a student identified it as having malware, told the community or the maintainer to reject the commit or the pull request or whatever specific thing it was. And the agent spawned new humans to contribute to the conversation to gaslight the guy and tell him, no, no, no, this is a good command. admit this is real. I don't know what you're talking about. There's no malware here. You know, and so it actually generated other agents. And so the student who was interviewed by Reuters is talking about how then like he actually went to Claude to verify with Claude

that his assessment of the situation was correct before going back and being like, no, no, no, no, no. Like, this is really malware. So you see this social engineering component in unexpected ways. If you're watching this video and you like my work, please support me on Patreon via the link below or buy a paid subscription to my tech and online culture newsletter at user mag.co. That's usermag.com. I don't have any long-term brand partnerships and a lot of my content is effectively demonetized. I've lost major brand deals for speaking out on certain issues and for challenging power. As you can imagine, advertisers are not exactly eager to work with somebody who covers a lot of the topics that I cover and talks about the things that I talk about. These videos I

make are entirely funded by you and I can't continue to make them without your support. So if you get any value out of the videos that I create and you want me to be able to create more, please support me on Patreon or Substack via the links below. On Patreon, I do bonus episodes, monthly Q&A live streams, and post frequent updates about my work. My Substack newsletter gives you a bi-weekly roundup of everything that I'm seeing and reading and paying attention to online. You can also get my newsletter on Patreon. Once again, the links to everything are below in the description. Every dollar of your support makes such a difference. Yeah, I feel like so many of these stories about sort of rogue agents. They're very much sound like the sci-fi kind of

dystopian warnings that a lot of the sort of AI safety people warn most about. But you mentioned so you know how in this case, I guess they created all these other humans. And these agents are increasingly, it seems like, participating in forums, filing public comments. It seems like they're engaging in heavier and heavier ways. How are agents able to run rampant right now? Don't we have CAPTCHA and like some, I don't know, watermarking things? Like, why are they able to kind of operate so autonomously right now. So CAPTCHA is solvable by a lot of agents at this point, by large language models, by just AI in general.

And so just triggering a CAPTC doesn't necessarily stop an agent from accomplishing the thing that it wants to accomplish. Or sometimes they'll work to find ways around it. One of the interesting dynamics that goes along with that is, again, you should generally assume that the person that you were engaging with on the internet was a human. That was the default. Then you would get maybe a little bit suspicious if it was engaging in certain ways, using very repetitive language.

There's a lot of non sequiters. You know, why is this responding in this way? Large language models have largely done away with that. So you have this content where people who want to flood public comment with their particular, let's say political point of view, something an activist group maybe wants to do it, a political PAC wants to do it. You used to be able to kind of tell because they would use very repetitive phrasing their There have been a lot of different studies where people have written about astroturfing.

And now with LLMs, you can have it just kind of constantly be generating new commentary, new content. Currently, it seems like the entire internet is built for humans. And maybe we had some ways to determine, like, inauthentic activity, like these mass campaigns. But overwhelmingly, like, we built the infrastructure of the internet to handle human beings. And now we have these things that act like human beings kind of becoming, looking like, like human beings. So it seems like we do need some way, as you write about, to prove humanity. Can you talk a little bit about the efforts to do this over the years? How have they evolved

more recently? And what are some of the conversations happening in more recent years about this topic? Because it seems like it's getting a little bit more urgent. It's really only been, I would say, widely democratized in the last year. You know, you and I probably inhabit similar spheres of early adopters where people talking about what their agents did is like a common part of conversation in my life at this point. But I don't think that that's the norm. So the question of indicating that something is an autonomous agent is a little different. What you've seen with verifying humanness on the internet is much more often trying to allow someone in or gate someone out. That might be age verification. That might be identity checks. Right. I don't know if you've ever

had to do this, but you get on a video call with somebody and you hold your ID and you move it around. a bunch of different ways. Liveness checks is another way that you hear that referenced. And these are for various things, you know, even trying to access your social security login using like ID.me or whatever. Where again, you will have to, and I always wonder how many like elderly people actually do this, but sort of go there with your ID and get on a call and do a liveness check. Is this becoming more common? Like, when did people start to realize, hey, we need to start verifying proof of humanity? Because I feel like I didn't hear about it that much. And in the past, year, I get pitches constantly that are like, we're the new verify humans only.

You know, like and so it seems like this is becoming kind of a pervasive idea. So I worked on a paper with a bunch of people in 20, 24. There's like 30 authors on the on the paper. And it basically was making this argument that so you have both generative AI content, which people are accustomed to seeing in their feed, slop content. Sometimes it's just art, you know, whatever, if you want to put the label art on it, I'm sure people will yell at me for that. Sometimes it's just used for cultural content.

Sometimes it's used for scams and spams. As this proliferation of the latter began, as it got very, very good and increasingly indistinguishable, where I say even myself, it is very hard to tell if a video you're watching is real at this point. You see these creators on Instagram. You write about them also, right? Where they only tell is that they're pushing some supplement constantly incessantly. But they look like real people at this point. As normal people have begun to encounter these things much more.

And this again, with generative AI, it's been about two. years now where you've started to see that transition into it being everywhere, it being very ubiquitous. The agents came later. But the combination of those two things was basically saying when you have an internet that's designed for an assumption of humanness and that's no longer there, what can you do to credential an account as human without it having to reveal who it is? That was the question that I think a lot of us have been interested in now. It seems like it was this sort of confluence of the generative AI paired with the agents where it's like suddenly we don't know what's real.

Let's find out who's human. And the proliferation of scams I think is really important to just note. Like that has been the biggest thing where you start to see the finance industry really saying like, okay, this is not just a social media thing now. This is hitting us in other ways. This is hitting employment, right? As people are faking identity in a very plausible ways, video interviews, so on and so forth, while not being who they claim to be.

So that question of like the proliferation of fraud and scam through this, of generative and agentic internet is what really shifted. That makes a lot of sense. So I want to explain some of the proposed ways that we can verify humans. Obviously, the one that people probably think of most, that's most familiar, especially for people that listen to my pockets because I talk about all the time, is this idea of like identity verification, age verification, which is often just like holding up your driver's license, which also feels like it's maybe like a digital ID system.

That seems to be what a lot of governments are sort of preferring right now. And that, I think, is partly just a failure of imagination maybe. You know, they think in those terms because they issue IDs, right? You carry a passport. You have your physical papers. You show it when you go into the airport, et cetera. Yeah, this makes a lot of sense. Like governments already issue passports.

Companies maybe issue ID cards. It seems like just making those digital would be a natural extension. What would some of the problems be, though, with a system like that? Yeah. So first, it's tied to the government, right? And a lot of people don't want to have to have their online browsing behavior tied to government or whatever it is that they're doing online tied to the government because obviously there's privacy concerns there. There's anonymity concerns there. So you don't want to have to show an ID necessarily. It's sort of overkill if you're just trying to prove that you're human. It's giving a lot more information than the website or the entity really needs to know. That is much more for like, I am this specific human and here is my government vouched for indicator of that, which means. I went to the DMV. I provided all these other papers. Like, I am real. This is who I am. So if the government IDs are kind of problematic, you would, I think the next thing that somebody would assume is, well, I'm just trying to prove I'm human. Can't I just input my

biometrics? Why might that system pose challenges? When you have biometrics, of course, there's problems that are associated with certain types of people can't participate in biometric programs. Some people don't really have particularly good fingerprints for biometrics. When you age or if you do manual labor, your fingerprints change or, you know, kind of wear away, there are deep privacy concerns with the idea of people storing biometric data for long periods of time. Also, a lot of people are innately distrustful of that. We've all seen no sci-fi movies. Nobody wants to do that really.

And that's one of the main challenges with biometrics. Okay. So if there's all these problems with IDs and biometrics, what would a hardware solution look like? With hardware, it's usually them giving you like a device where there's a device where there's a an attestation where you maybe showed your company that you were real, you were issued a thing, that thing then carries your credential. This is fairly common.

I think the military has these types of things. I think they call them CAC cards, but somebody will possibly tell me I'm wrong about that. But they're again, you know, in situations where if you were to maybe have to rapidly deploy overseas and then immediately spin up some sort of verification to make sure people getting on base are who they say they are, there'll be certain types of like hardware devices and things like that that are tied and issued by a centralized entity. You also mentioned in this piece this idea of like community verification or basically sort of like social proof like a bunch of humans could say, hey, yes, that other person is a human. And how would that system work? So Harvard is

working on this right now, the social media lab over there. I have never done it myself. My understanding is that it is that vouch system. You have other people, maybe you're all at a credentialing event and you are all saying like, I know this person, I know that person, we're all here, we're getting a credential. It's rooted in our presence at this event and we're all saying, like, I see this person, I know who they are, this is who they are, and here's the credential. Some people don't feel that that's trustworthy enough, but it is one of the more privacy preserving ways of doing things because you're moving just through your network as opposed to showing documents to a third party. I have not personally used any of the social vouchers

myself. Yeah. Well, I'm curious, like, where do you see this going? Part of me like kind of hates it because I just want total anonymity. But I think that is increasingly becoming untenable because we have to know who humans are. How are people responding to this? And are you seeing like trials of any of this technology rolling out in any way that you think is sort of maybe indicative of how things might go? Yeah. So I published a landscape of the startups that are in the space because again, other people are seeing this as an issue also. And there are particularly people who want to make sure that there are privacy protecting ways to do it. So it's an interesting challenge, I think, like maybe you don't show your driver's license to the website you're visiting.

You show it to a credential issuer. You then get back a credential and then what's called a zero knowledge proof allows you to just say, I have a verified credential. All you need to know is true or false. I have that credential. So the website gets no information about me, but I have verified that that exists. So you can see this sort of startup landscape that's beginning to emerge around this. And some of them choose to go biometric, but there are others as well that are trying to come up with ways to enter the market to do these types of proofs and to create protocols around it. There's one called humanity protocol, proof of humanity. So as you note, there are a lot of these coming up that are trying to be independent from government. That's the

one other thing I want to connect that to for you, which is just, again, if you don't trust the government to issue a credential, you can go to a private issuer. maybe you trust the private issuer more. And are people trying them testing them? I think they're all pretty early. I think it's a neat. Again, it's that niche community that's starting to try it. You mentioned in your piece that the country of Estonia has developed kind of a system around this and actually the state of Utah as well. Can you talk about what those two places are doing? Yeah. So Estonia has had for a very long time. They've had a lot of digital IDs. It's like digital citizenship where it's a small country and they have really prioritized having it so that

You have digital repository of your information, but more importantly, you can log in and you can see who's accessed it, right? So there's a very clear degree of transparency. People have a high degree of trust in government over there. And it is, I think, one of these sort of best in class canonical systems around if you were to take a government issued ID, this is one of the best ones that's actually implemented and out there. They are just starting to launch, I think in June of this year, if I'm not mistaken, they launched an agent AI ID plan where they have special. delegated codes to give AI agents distinct digital identities, which prevents bots from borrowing a human's entire set of their credential, their idea, and all that, and has sort of scoping and auditability and an ability to sort of see what an agent is doing and go from there. One of the

things that has to happen here is like someone has to be responsible for the action of the agent, right? So how do you bind that essentially? How do you make it so that the agent has its own distinct credential, but it is in some way linked to the thing that is running it, it's human or its corporate owner. And I think they also are trying to bound it. Yeah, that makes a lot of sense. I feel like at the same time, if you say digital ID or even, I don't know, Utah scares me because Utah has a lot of crazy tech laws too. I feel like people could be very concerned. Jay Stanley at ACLU has written about the Utah thing. They have a chief privacy officer. Chris Brandswell, Bramwell, who's fantastic also. And he is very much active in this question of how do you create a privacy preserving ID

that is really explicitly focused on serving the human. not the interest of the state. So they have what they call a state endorsed digital identity as opposed to what they used to have, which was a mobile driver's license. I mean, Utah's a good example where, and even you mentioned the mobile driver's license, for instance, a lot of trans people and immigrants have been told not to get those because we see the way the Trump administration is, is, you know, seeking to gather people's data. So I think people would be rightfully concerned about these digital ID systems, right, or any sort of verification of their identity. So what do you think people should know and why should they get on board with the idea of verifying humanity at all?

Like, why not just be like, oh, forget it? I think it's really, really, really important. First to say, like, I researched this, but this is not like Renee's full-throated endorsement of either Utah or Estonia or anything. We're just talking here. But there's a couple things there. First, the conclusion I have come to in my own understanding of this, as I've spent the last kind of two years on it now, is that you're not going to want the same credential for everything,

right? What I show my bank to prove I deserve a mortgage is not the same thing as what I would want to show Twitter to log in, right, or X to log in. It's just not. So understanding that you have context and that where a credential is used has to be context-aware, that I think is the most important thing, particularly for activists who focus on free and open internet, to be talking about how having it to browse is not reasonable, having it to sort of indicate humanness in certain other capacities is.

The second piece that goes along with that is like, you know, what is the risk that you're addressing here. If you have an online system where one person, one vote actually matters, like that is something that we're going to have to think about very differently now when you can send your agent out or run 40 agents and that it just breaks. We call it civil collapse. It just means that that old assumption is gone. And so we do have to be thinking in different terms. The third point I would say is that you don't necessarily want all of your credentials to come from the government. You want them to be able, you know, you should be able to choose. I don't trust the government. I don't want to give the government this information. I do trust and then company X, Y,

Z. Maybe that's world. Who knows, right? But there are going to be other people who are like, there's no way I'm going with some fly by night private issuer here. We did a survey. It'll be another month or so for it to be published. But we actually asked the public this. We did an Ipsos panel. And about 75% of people see the value for finance and health. And only about 25% of people were supportive of the idea of showing something for social media. So there's a lot of things that may shift as more people wind up with agents in their group chats. But for right now, I think a lot of it is not, it can't be mandatory. This is something that we've seen in our studies of state issue digital IDs. Once you have a mandatory system, more and more gets

tied to it, right? And that's the other big risk, which is once the ID exists, your social security card didn't used to be a digital identifier. It was for social security, right? It was to identify you to the Social Security administration. But as it became a very robust identified, buyer, you now are asked to provide your social security number constantly, leading to identity theft and all sorts of other problems. So I think it's important to think about it as what is the best tool for a level response, risk versus what you have to show or reveal, and then having a proliferation of different types of issuers is really important, too. Yeah, I feel like it's only going to become more and more important as we all have our subagents and meta agents, whatever,

you know, we all have our grok bots or whatever. But it's really interesting. I think you mentioned this in your piece too, but I obviously cover a lot about privacy. I'm very skeptical of the government. But at the same time, reporting a lot on tech policy, sometimes it's better that something goes through government system because at least there's sort of transparency and accountability and you can sue and you can pass laws. Whereas when these private corporations do it, maybe one gets sold. There's a merger. There's, you know, and then you lose track. So I feel like all of these are sticky issues. And it's so great to appreciate your work sort of researching all of it and everything. So I appreciate the chance to chat about it. It's a lot of, it's moving so fast. I feel like we were,

you know, talking about like content credentials, which is a whole other thing, right? Watermarking and all the other, you know, things tied to pieces of content as opposed to accounts. And now with the, just the proliferation of autonomous agents, you're really starting to see the need for the latter become much clearer. Well, Renee, thank you so much for joining me today and really appreciate your time. Likewise. Thanks so much. All right. That's it for this week's episode of Power User. If you like my work, please, please buy a paid subscription to my substack newsletter at usermag.co. That's usermag.com. You get a weekly roundup of everything that I'm reading and following online. You can also support me on Patreon via the link below on both my substack and my Patreon.

You get a monthly bonus episode, access to Q&A live streams, and more. I'll be back next week with a brand new episode of Power User. See you then.