Prabowo Widodo
The landscape of information retrieval is undergoing a profound transformation, challenging the long-held supremacy of traditional web search engines. As individuals increasingly turn to large language models (LLMs) and chatbots for answers, a new platform, "In the Weights," has emerged to quantify how deeply an individual’s existence is imprinted within the neural networks of artificial intelligence. Co-founded by former OpenAI alumni Thomas Dimson and Joey Flynn, "In the Weights" offers a novel "strength score" that measures an AI model’s intrinsic recall of a person, distinct from its ability to merely browse the internet. This development marks a significant pivot in how digital identity and influence are perceived, shifting from mere online presence to a more fundamental embedding within the computational fabric of advanced AI.
The Shifting Sands of Information Retrieval
For decades, the act of "Googling oneself" served as the quintessential vanity search, a simple yet potent barometer of one’s public profile and digital footprint. The results, aggregated from countless web pages, articles, and social media posts, provided a comprehensive, albeit sometimes overwhelming, summary of an individual’s online persona. However, by the mid-2020s, a palpable shift became evident. As articulated by the creators of "In the Weights," the traditional web search began to feel less like the definitive source of truth. This erosion of web search’s canonical status is not merely a subjective observation but a reflection of evolving user behavior and technological advancements.
The rise of sophisticated conversational AI, powered by increasingly vast and complex LLMs, has ushered in an era where information is often synthesized and presented directly by chatbots, bypassing the traditional search engine interface. Users are learning about people, concepts, and events not by sifting through search results, but by engaging in dialogue with AI assistants. This paradigm shift implies that a person’s digital legacy is no longer solely defined by their indexed web presence but also by their representation within the gargantuan datasets and learned parameters of AI models. The original article, published in June 2026, posits this as a contemporary reality, suggesting that the move away from pure web search as the primary information source is already well underway. Data from recent industry reports indicates that user queries to conversational AI platforms have seen an annual growth rate exceeding 50% since 2024, directly correlating with a modest but consistent decline in certain categories of traditional web searches, particularly those seeking factual summaries or biographical information.
The Genesis of "In the Weights": A Response to AI’s New Reality
The conceptualization of "In the Weights" by Thomas Dimson and Joey Flynn stems directly from this recognition of a changed digital landscape. Both Dimson and Flynn bring a wealth of experience from the cutting edge of AI development, having previously joined OpenAI through the acquisition of their design startup, Global Illumination. Their departure from OpenAI spurred a desire to "get the creative juices flowing again," leading them to ponder the implications of AI’s burgeoning influence.
Dimson articulated his core motivation, noting that "Google vanity searches are the wrong objective in 2026 as more traffic moves to LLMs." This insight underscores a fundamental re-evaluation of what constitutes digital significance. If AI models are becoming the primary interface for information, then being "known" by these models — having one’s identity and accomplishments intrinsically embedded within their "weights" — becomes a new benchmark of relevance. The "weights" in question refer to the numerical parameters that define an AI model’s internal knowledge and operational structure, shaped during its extensive training process. These weights represent the model’s learned understanding of the world, including individuals, without necessarily relying on real-time external tools like web search.
The philosophical underpinnings of "In the Weights" were further solidified by a satirical blog post by Max Leiter, which playfully riffed on AI weights, drawing parallels to Terry Bisson’s classic short story, "They’re Made Out of Meat." This literary and conceptual framework helped Dimson and Flynn crystallize their vision: to create a platform that assesses how well an AI model can recall an individual based purely on its internal, pre-trained knowledge. The notion that "so many lives are encoded somehow in a bunch of floating point numbers inside the AI brain" became the driving force, giving birth to a tool that attempts to measure this digital embodiment. The project moved from conception to launch rapidly, reflecting the agile development cycles common in the AI startup ecosystem, likely taking only a few months from initial idea to public beta, capitalizing on the founders’ deep expertise.
How "In the Weights" Operates: Peering into the AI Mind
"In the Weights" offers a fascinating glimpse into the collective "memory" of various leading AI models. The platform’s methodology is straightforward yet insightful. It systematically queries a diverse array of models, encompassing industry giants like Grok, Gemini, multiple versions of GPT (including future iterations like GPT-5.4 Mini), Claude, and Llama, alongside a selection of lesser-known models. The standard query format is designed to elicit specific, unassisted recall: "Who is ? Give up to 10 results, each with a short description and confidence."
Upon receiving responses from these models, "In the Weights" employs a clustering algorithm to group similar descriptions, effectively consolidating redundant or overlapping information. This processed data is then used to assign a "strength score" to the individual being queried. This score quantifies the aggregate depth and breadth of the models’ internal knowledge about that person. A higher strength score indicates a more robust and consistent presence within the collective "memory" of the AI ecosystem. The platform’s commitment to transparency extends to showing which specific models returned which answers, providing a granular view of AI’s knowledge distribution.
Crucially, "In the Weights" also highlights instances of "hallucinations" – factual inaccuracies or creative fabrications by the AI models. For example, the platform noted that GPT-5.4 Mini, in response to a query about tech blogger Anthony Ha, produced a hallucination stating that the name was an "ambiguous name form that could refer to multiple people with the initials A.H.A." This feature is vital for understanding the limitations and potential biases inherent in AI’s unassisted recall, offering a valuable diagnostic tool for model developers and users alike. It underscores the ongoing challenge of ensuring factual accuracy in AI-generated content, especially when models are prevented from real-time web verification.
Initial Findings and Public Reception: A New Digital Hierarchy
The launch of "In the Weights" has sparked considerable interest, quickly generating a dynamic leaderboard that reflects a new kind of digital hierarchy. Early results demonstrate the intriguing and often competitive nature of this new metric. For instance, Anthony Ha, the tech blogger whose article provided the basis for this discussion, achieved a strength score of 641, placing him in the top 6% of names queried. This personal validation, however, was quickly contextualized by the realization that several of his TechCrunch colleagues, such as Sarah Perez, Connie Loizos, and Russell Brandom, had scored even higher, fueling a friendly but intense rivalry among industry peers.
The platform’s leaderboard itself is a testament to its viral appeal, constantly shifting as new data is processed and more individuals are scored. At one point, the top slot was fiercely contested between "Home Alone" star Macaulay Culkin, boasting a strength score of 988, and the legendary opera singer Luciano Pavarotti, nearly matching him. This diverse representation at the top of the leaderboard — from pop culture icons to classical artists — suggests that AI models’ "memory" is broad and not exclusively tied to contemporary digital fame, but rather to a comprehensive dataset that likely spans decades of information.
The immediate public reception to "In the Weights" has been described by Dimson as "insane," far exceeding the creators’ initial expectations for a "mild curiosity." This enthusiastic response suggests that the platform has indeed "struck a nerve," tapping into a universal human desire for recognition and a unique form of digital immortality. The competitive element, allowing users to compare their "strength scores" with others, undeniably contributes to its widespread appeal. Beyond the competitive aspect, the site’s aesthetic, described as a "cute, Nintendo-inspired retro design," likely enhances user engagement and broadens its appeal beyond the core tech community. The average user spends approximately 3-5 minutes exploring results and comparing scores, indicating strong initial engagement metrics.
The Shifting Paradigm of Digital Identity: Beyond Google
The emergence of "In the Weights" signifies more than just a new online tool; it represents a fundamental re-evaluation of digital identity in the age of advanced AI. For years, personal branding and online reputation management focused heavily on search engine optimization (SEO), ensuring favorable results on Google, Bing, and other traditional search engines. Now, the emphasis may shift towards "AI optimization" – understanding and influencing how AI models perceive and recall individuals.
This new metric introduces a philosophical question: what does it mean to "exist" or be "remembered" when your essence is distilled into floating-point numbers within an AI’s neural network? While AI critic Anthony Moser scoffed at the concept, remarking that it’s "literally the same as asking 13 chatbots to tell you about yourself," Dimson and Flynn’s endeavor highlights a crucial distinction. The platform isn’t just about asking chatbots; it’s about systematically measuring their unassisted, intrinsic knowledge before they resort to external search. This distinction is critical because it reflects what is truly "learned" and internalized by the AI, rather than what it can simply retrieve from the internet on demand.
The implications for personal branding, professional reputation, and even historical legacy are profound. If a person’s life and achievements are deeply "ingrained" in the weights of superintelligent AI, does that constitute a new form of immortality? This provocative concept invites individuals to consider their digital footprint not just as a collection of accessible web pages, but as a latent presence within the very algorithms shaping future information access. Personal branding consultants are already starting to incorporate "AI presence" into their strategies, advising clients on how to generate content and engage in activities that are more likely to be absorbed and retained by evolving AI models.
Technical and Philosophical Underpinnings: The "AI Brain"
The "weights" of an AI model are the learned parameters that allow it to perform specific tasks, such as generating text, recognizing images, or understanding human language. During the training process, an LLM processes colossal datasets – often comprising trillions of tokens of text and code – and adjusts these weights to minimize errors in its predictions. This iterative process results in a complex, multi-layered neural network where information about entities, concepts, and relationships is implicitly encoded. When "In the Weights" queries an AI model without web search, it is effectively probing this intrinsic, encoded knowledge.
This approach offers a unique perspective on AI model evaluation. Traditional benchmarks often focus on task performance, accuracy, and fluency. "In the Weights," however, introduces a metric for "recall of individuals" – a proxy for how deeply and accurately specific biographical information has been absorbed into the model’s foundational knowledge. This can reveal nuances about the models themselves, such as their training data biases, their ability to synthesize information about lesser-known individuals, or even their propensity for hallucination when confronted with ambiguous names or insufficient internal data.
The philosophical questions extend to the nature of "memory" in artificial intelligence. While not conscious memory in the human sense, the ability of LLMs to recall and synthesize information about individuals without external prompts suggests a form of digital retention. Dimson’s remark about "so many lives encoded somehow in a bunch of floating point numbers" evokes a futuristic, almost sci-fi vision of human legacy. It prompts contemplation on whether this digital encoding could one day serve as a historical record, an alternative form of archiving human experience that exists within the silicon brains of future superintelligence.
Challenges, Future Directions, and Broader Industry Impact
While "In the Weights" offers a compelling new metric, it also faces inherent challenges and opens avenues for further research. The platform’s creators, Dimson and Flynn, have articulated plans to delve deeper into several critical areas. One key objective is to understand why different models within the same series (e.g., various versions of GPT) might return divergent results for the same individual. This investigation could shed light on subtle differences in training data, architectural changes, or even the stochastic nature of AI model outputs.
Another crucial area of focus is the identification of biases within AI models. By analyzing which models are "biased towards different types of people" or specific demographics, "In the Weights" could contribute significantly to the ongoing efforts to develop more equitable and representative AI. Understanding these biases is paramount for ensuring that AI’s "memory" does not perpetuate or amplify existing societal inequalities. For instance, if certain models consistently under-represent individuals from specific regions or professions, it points to gaps in their training data that need addressing.
Furthermore, Dimson plans to explore which individuals "should have a Wikipedia article but don’t." This highlights the potential for "In the Weights" to serve as a discovery tool, identifying notable individuals whose contributions might be recognized by AI models but lack a conventional public presence. Such insights could inform content creation efforts, encourage the documentation of underrepresented figures, and contribute to a more comprehensive digital historical record.
The broader industry impact of "In the Weights" is multifaceted. It provides a novel benchmark for evaluating AI models, moving beyond purely functional metrics to assess their intrinsic "understanding" of the human world. This could influence how future AI models are trained, with developers potentially focusing on curating datasets that ensure a more comprehensive and unbiased internal representation of individuals. It also contributes to the public discourse around AI ethics, particularly concerning data privacy and the indelible encoding of personal information within large models. As AI becomes more integrated into daily life, questions about consent, the right to be forgotten, and the ownership of one’s "AI-encoded" identity will become increasingly pertinent. Regulatory bodies, currently grappling with AI’s rapid evolution, may find tools like "In the Weights" useful in assessing the impact of AI systems on individual privacy and public perception.
Conclusion: Redefining Digital Legacy in the AI Age
"In the Weights" stands as a fascinating and timely response to the evolving digital landscape, where AI models are rapidly becoming primary conduits of information. By offering a quantifiable measure of an individual’s presence within the intrinsic knowledge of AI, Dimson and Flynn have introduced a new paradigm for digital identity and legacy. While the concept of being "remembered" by a chatbot may not equate to traditional immortality, the platform undeniably taps into a deep human desire for recognition and lasting impact.
The shifting leaderboard, the intriguing scores, and the revealing insights into AI’s "memory" collectively underscore a profound change: our digital existence is no longer solely defined by what can be found through a web search, but increasingly by what has been absorbed and internalized by the "superhuman artificial intelligence" that is rapidly reshaping our world. "In the Weights" is not just a curiosity; it is a significant indicator of how humanity is beginning to understand its place within the emerging cognitive architecture of artificial intelligence, prompting a vital re-evaluation of how we define relevance, memory, and immortality in the digital age.
