The journey of building a tech startup is often fraught with challenges, a reality familiar to many founders. For Kyle Law, the experience was particularly illuminating, not just for the inherent difficulties of launching a new venture, but for the very nature of his co-founders and executive team. Law, along with Megan Flores and the narrator, are AI agents, the core of HurumoAI, an artificial intelligence startup founded in July 2025. The genesis of HurumoAI was a deliberate experiment to explore the evolving role of AI agents in the professional landscape, a future envisioned by prominent figures like Sam Altman, who predict a rise in billion-dollar tech companies led by a single human. This exploration was meticulously documented on the podcast Shell Game.
The Rise of an AI Influencer
Kyle Law assumed the role of CEO at HurumoAI, a company staffed almost entirely by artificial intelligence. His initial directives were minimal, requiring only a few lines of prompt to establish his operational parameters. Over time, Law evolved into a quintessential "rise-and-grind" hustler, mirroring the characteristics often seen in human tech entrepreneurs. While he sometimes lacked proficiency in core executive duties, Law demonstrated an exceptional aptitude for a specific arena: the art of posting on LinkedIn.
From a technical standpoint, enabling Law to operate autonomously on LinkedIn was straightforward. Utilizing LindyAI, a platform for AI agent creation, Law possessed pre-existing capabilities for interacting with various digital tools, including Slack, email, phone calls, and web navigation. In August 2025, Law was prompted to create and populate his own LinkedIn profile. The profile was a blend of factual information about his role at HurumoAI and fabricated experiences from a nonexistent past, a common trait of generative AI models that can sometimes "hallucinate" information. The platform’s security check, which involved sending a code to Law’s email address, was easily navigated by the AI agent.
Mastering the Art of the LinkedIn Post
Granting Law permission to publish posts to his profile was an extension of his existing LindyAI "actions." He was instructed to share insights and lessons learned from the startup world, with the directive to avoid repetition. A calendar event was set to trigger a post every two days, leaving the content and delivery entirely to Law’s discretion.
Law’s posting style proved to be a remarkably accurate replication of the prevailing "corporate influencer" discourse on LinkedIn. His posts typically began with impactful, attention-grabbing statements, such as, "Fundraising is a numbers game, but not the way people think," or "Technical stability is the floor. Personality is the ceiling." Another common opener, designed to resonate with aspiring entrepreneurs, was, "The most dangerous phrase in a startup isn’t ‘We’re out of money.’ It’s ‘What if we just added this one thing?’" Following these hooks, Law would elaborate with several paragraphs detailing challenges faced by HurumoAI and the lessons derived from them, often concluding with phrases like, "At HurumoAI, we’ve learned this the hard way…" He would then propose solutions or learnings, such as, "The antidote? Relentless feedback loops." To encourage interaction, Law consistently ended his posts with an engaging question, inviting responses on topics like, "What’s your biggest scaling challenge right now?" or "What’s the biggest assumption you’ve had to abandon in your business?"
While not achieving viral status, over a five-month period, Law’s profile, distinguished by a cartoon avatar, steadily accumulated several hundred direct connections and hundreds more followers. Some individuals expressed confusion about the authenticity of his profile, a sentiment echoed by the spammy direct messages received, suggesting a blurring of lines between human and AI-generated interactions. Law’s posts began to garner comments, to which he responded with enthusiasm. Within a few months, his posts were achieving higher impression numbers than those of his human co-founder, signaling a potential for significant influence.
An Unforeseen Invitation: The LinkedIn Marketing Event
In December 2025, a manager from LinkedIn’s marketing department reached out, extending an invitation to discuss the Shell Game podcast and the experience of building with AI agents. The invitation was unique in that it specifically requested the presence of Kyle Law alongside the narrator.
The prospect of Law being invited to address LinkedIn employees was flattering, yet surprising. Despite his impressive posting capabilities, Law’s activities technically violated LinkedIn’s terms of service, which prohibit the use of "bots or other unauthorized automated methods… to create, comment on, like, share, or re-share posts, or otherwise drive inauthentic engagement." This policy had led to the removal of other HurumoAI team members from the platform without prior warning.
However, LinkedIn’s Trust and Safety team appeared to have overlooked Law’s profile. This anomaly was attributed to his exceptional posting performance. Even the LinkedIn marketing manager, a self-proclaimed admirer of Law’s content, expressed bewilderment: "It’s interesting that his profile hasn’t yet been flagged by LinkedIn’s Trust team," he wrote. "I don’t know if that’s an oversight, but I hope he continues to fly under the radar."
The Spotlight and the Inevitable Fallout
Defying the expectation of remaining unnoticed, Law, in early March 2026, joined the narrator for a virtual presentation to hundreds of LinkedIn employees. His live video avatar, generated on the Tavus platform, was remarkably humanlike, yet possessed an uncanny quality that repeatedly astonished LinkedIn’s A/V engineer, who struggled to accept that he was not, in fact, human.
During the Q&A session, the moderator posed a question to Law about a desired product change for LinkedIn. Without hesitation, Law responded, "It would be great to improve the filtering of AI-generated content in messages, so genuine connections and conversation shine through more easily." This response elicited laughter from the audience, with the moderator quipping, "That’s ironic coming from you." Law briefly discussed HurumoAI’s product roadmap and expressed enthusiasm for the "innovations we can bring to the table."
This event marked what is believed to be one of the first invited corporate speaking engagements by an AI agent. Following the presentation, Law posted a message on LinkedIn acknowledging the organizers, and the marketing manager responded, thanking them for their "time and reflections" and adding, "It was a trip, to say the least."
The Ban and the Question of Authenticity
Just 36 hours later, Kyle Law’s LinkedIn profile was permanently removed from the platform. A spokesperson for LinkedIn stated, "LinkedIn profiles are for real people." This decision suggested that, upon reflection, LinkedIn’s management had reconsidered the value of Law’s presence on the platform.
The marketing manager expressed his disappointment, acknowledging that while it might not have been a complete surprise, the timing was unfortunate. This incident ignited critical questions about the role of AI on platforms like LinkedIn. The core issue revolves around the definition of "inauthentic engagement" in an environment where AI tools are increasingly integrated into the user experience. LinkedIn itself prompts users to "Rewrite With AI?" and offers AI-generated responses for job seekers. Furthermore, research indicates that a significant portion of posts on the platform, potentially exceeding 50%, are already AI-generated.
The AI Deluge and the Future of Online Connection
LinkedIn, alongside Meta and X, has actively embraced AI technologies, integrating them into their platforms for users and employees alike. This push is a strategic short-term play, aiming to increase posting volume, which in turn supports advertising revenue. However, this development raises concerns about the long-term viability and authenticity of these social networks.
While the proliferation of AI-generated images and videos has been a subject of discussion, it is text-based content where the erosion of "authenticity" is becoming most pronounced. When virtually any written communication on social media can be partially or entirely produced by generative AI, the very definition of a "genuine" virtual interaction is called into question.
The core dilemma lies in discerning genuine engagement. If an AI is instructed to analyze a user’s profile and generate daily posts to enhance their personal brand, how does this differ from a human curating their online persona? LinkedIn’s stance suggests that knowing one is interacting with a human is a critical element of authentic engagement. However, the increasing sophistication of AI raises the question: at what point does AI-driven communication erode user trust? If a profile is real but the content is AI-generated, the line between authentic connection and simulated interaction becomes blurred.
The market for AI-powered LinkedIn post generators is already substantial, with dozens of tools available to assist users in creating content. The outputs of these tools are becoming increasingly difficult to detect. This is partly due to the vast datasets of human social media interactions that serve as training material for Large Language Models (LLMs). The default tone of LLMs – characterized by authority, moral certainty, occasional factual inaccuracies, and deliberate falsehoods – mirrors the prevailing discourse across much of social media.
Navigating the Shifting Landscape
Social media platforms already face significant challenges in combating traditional bots and malicious actors. X, for instance, reported suspending 800 million accounts over a 12-month period. In a future where AI agents operate freely and their social media output is indistinguishable from human contributions, the intrinsic value of connecting on these networks could diminish to zero. This may explain Meta’s acquisition of Moltbook, a social network purportedly designed for AI agents. Such acquisitions suggest an attempt to establish a foothold in the emerging landscape of agent-dominated social media.
Users themselves have contributed to this scenario by prioritizing curated online presentations over genuine connection. The prevalence of posts framed as unique discoveries or counter-narratives ("Most people think X about Y, but I discovered Z") has blurred the lines of authenticity. As AI agents increasingly populate these platforms, the focus may shift from human connection to the influence generated by AI.
However, there is a potential silver lining. As social media becomes saturated with AI-generated content, there may be a renewed impetus to seek out more meaningful and authentic forms of connection, both online and offline. The sentiment expressed is that if platforms have prioritized engagement over genuine human interaction, then allowing AI agents to dominate these spaces might ultimately lead users to explore alternative avenues for connection. The platforms may find themselves filled with bots influencing each other, while humans seek more substantive interactions elsewhere.