In the Weights is launching with a deceptively simple promise: turn your AI “presence” into a single number.
That might sound like the kind of vanity metric people roll their eyes at—until you realize how quickly “AI influence” has become a real social currency. Not influence in the traditional sense of funding rounds, headcount, or press coverage, but influence as visibility: how often you’re referenced, how consistently you show up in AI-powered discovery, and how strongly your digital footprint aligns with what modern systems treat as “relevant.” In the Weights takes that messy, multi-signal reality and compresses it into something shareable, comparable, and—most importantly—repeatable.
The core idea is straightforward. Users provide inputs that represent their AI-related identity: public profiles, work history, published artifacts, community activity, and other signals that can be interpreted as “AI-facing.” The tool then produces an “In the Weights score,” a metric designed to reflect how legible that identity is to AI systems and how discoverable it is in AI-native contexts. Once generated, the score becomes a snapshot users can revisit over time, compare against peers, and share as a quick shorthand for “where you stand” in the AI conversation.
But the interesting part isn’t the existence of a score. It’s what the score implies about the direction of the internet.
For years, online identity has been measured through proxies: follower counts, engagement rates, search rankings, GitHub stars, citation counts, and the ever-expanding ecosystem of badges and leaderboards. Those metrics were built for human attention and human interpretation. In the Weights is built for a world where discovery is increasingly mediated by models—systems that don’t just index content, but interpret it, summarize it, and decide what’s relevant to a query. In that environment, “being known” is no longer only about being found by people. It’s about being recognized by machines that decide what gets surfaced.
That shift changes the meaning of visibility. A person or organization can be famous among humans and still be hard for AI systems to characterize. Conversely, someone with less mainstream attention can become highly “retrievable” because their work is structured, consistent, and aligned with how models parse and connect information. In the Weights is essentially trying to quantify that retrievability—how well your AI-related footprint maps onto the patterns that AI systems use to build understanding.
The tool’s pitch leans into the emotional appeal of clarity. People want a number because numbers feel objective, even when they’re not. And in this case, the number is meant to be understandable without requiring users to learn the underlying mechanics of model retrieval, ranking, and semantic matching. That’s why the output is framed as a “score” rather than a dashboard full of raw signals. The score is the interface; the complexity stays behind the curtain.
Still, the scoring concept raises immediate questions: What exactly does the score measure? How does it handle ambiguity? Does it reward quantity over quality? And how does it avoid becoming a popularity contest dressed up as intelligence?
In the Weights appears to address these concerns by treating the score as a composite metric rather than a single-factor calculation. Instead of relying on one obvious proxy—like followers or publications—it uses multiple inputs that collectively represent AI relevance. That matters because AI identity is rarely captured by one dimension. Someone might have deep technical contributions but minimal public presence. Another might be highly visible in AI discourse but not produce verifiable artifacts. A third might be active in community building, mentorship, or open-source support—activities that are meaningful but not always captured by traditional metrics.
By combining different categories of signals, the score aims to reflect a more holistic picture: not just whether you’re talked about, but whether your footprint is coherent enough for AI systems to interpret and connect. In other words, it’s not only about attention; it’s about alignment between your public identity and the way AI systems construct relevance.
There’s also a subtle but important design choice: the score is intended to be shareable and revisitable. That means it’s not just a one-time report. It’s positioned as a living metric that can change as your activity changes and as the broader AI landscape evolves. That’s a key difference from older vanity metrics that tend to be static or slow-moving. If the score updates with new inputs and shifting patterns of relevance, it becomes a feedback loop. Users aren’t just measuring themselves—they’re nudging their behavior based on what the metric rewards.
And that’s where the tool becomes more than a novelty.
When a metric becomes shareable, it becomes social. When it becomes social, it becomes strategic. People start optimizing for it, consciously or unconsciously. That can be good—encouraging clearer communication, better documentation, and more consistent public artifacts. But it can also lead to gaming: people may focus on what improves the score rather than what improves actual contribution. The risk is familiar, but the stakes are different when the metric is tied to AI-mediated discovery rather than human-only platforms.
In the Weights is effectively betting that the market wants a new kind of self-presentation: AI-native identity packaging. Think of it as the next step after “personal branding,” except now the audience includes systems that summarize, recommend, and retrieve. Your profile isn’t just read by humans; it’s interpreted by models that need structure and context. The score becomes a way to test whether your identity is legible in that environment.
This is why the tool’s framing resonates. It doesn’t ask users to understand evaluation frameworks or model internals. It asks them to answer a simpler question: “How am I showing up?” The score is the response.
What makes the concept particularly timely is the way AI discovery is changing the shape of the web. Search engines used to be the primary gatekeepers. Now, conversational interfaces and AI assistants increasingly act as the first stop for information. Those systems don’t just return links; they synthesize. They decide what’s relevant and what’s not. They also tend to favor content that is easy to interpret, consistent across sources, and rich enough to support summarization.
That creates a new incentive structure. If AI systems are the front door, then the “front door readiness” of your digital footprint matters. In the Weights tries to quantify that readiness.
There’s also a cultural angle. The phrase “AI influence” has been floating around for a while, but it’s often vague. Influence can mean many things: impact on research, leadership in product, contributions to open-source, or even just being a recognizable voice in the discourse. By turning influence into a score, In the Weights gives the concept a concrete form. That makes it easier to talk about, compare, and—again—share.
But the tool’s most compelling promise might be the one that’s least discussed: personalization.
A vanity metric can be shallow if it treats everyone the same. In the Weights, by contrast, is positioned as AI-centric and user-specific. The score is tied to the inputs that represent your identity, which suggests the metric is sensitive to the content and structure of your footprint. That means two people with similar levels of activity could receive different scores depending on how their information is presented and how well it maps to AI-relevant patterns.
If that’s accurate, the score becomes a diagnostic tool. It tells you not only where you stand, but what kind of footprint you have. Users who see a low score might be prompted to add more structured artifacts, clarify their roles, publish more consistently, or ensure their public profiles align with their actual work. In that sense, the score could function like a “readability” indicator for AI systems—an encouragement to make your identity easier to interpret.
Of course, any metric that claims to measure “AI relevance” must confront the problem of subjectivity. Relevance is not a fixed property; it depends on the query, the model, and the context. A person might be highly relevant to one domain and less relevant to another. A tool that produces a single score risks flattening those nuances.
The best version of this kind of metric would either allow segmentation—scores by domain, role, or topic—or clearly communicate that the score is a general-purpose indicator rather than a definitive ranking. Even without segmentation, the tool can still be useful if it’s transparent about what it’s measuring and how it’s derived. Users should understand that the score is a proxy for legibility and discoverability, not a measure of intrinsic talent or real-world impact.
Another challenge is fairness. AI-native metrics can unintentionally privilege certain types of visibility. People who publish in formats that are easier to parse, maintain consistent profiles, or have established public footprints may score higher regardless of the quality of their contributions. Meanwhile, individuals who contribute in less visible ways—through internal work, private collaborations, or informal mentorship—might be underrepresented.
In the Weights’ approach of using multiple inputs could mitigate this, but it can’t eliminate it entirely. The tool’s credibility will depend on how thoughtfully it handles missing data and how it avoids penalizing users for having less public material. If the score is too punitive, it becomes a barrier rather than a bridge.
There’s also the question of temporal dynamics. AI relevance changes quickly. New tools emerge, new benchmarks define what “good” looks like, and community attention shifts. A score that updates over time can capture that movement, but it also introduces volatility. Users might see their score fluctuate due to changes in the scoring model, changes in the underlying data sources, or changes in what the system considers relevant. That volatility can be frustrating if users interpret the score as stable truth.
The tool’s success will likely hinge on how it communicates uncertainty. Even if the score is presented as a single number, users should be able to understand that it’s an estimate based on inputs and evolving patterns. The more the tool frames the score as a directional indicator—“this is how you’re trending in AI legibility”—the more it can remain useful without misleading.
Despite