The Dating App Hell is Coming to The Job Market

J'ai écris cette semaine.

Mercor, co-founded by CEO Brendan Foody, is pioneering a new era in recruitment by leveraging AI to predict job performance more accurately than traditional human methods. Their systems automate processes like resume screening and interviewing, enabling them to identify highly capable individuals, often globally dispersed, for demanding roles, particularly training advanced AI models for top labs. This approach moves beyond simple credential matching, aiming to understand and quantify the potential for high performance in complex knowledge work.

Through this intensive work of evaluating candidates and tracking their subsequent performance, Mercor and others in the field are observing a distinct pattern emerge in the value generated by knowledge workers: a power law distribution. This means that, much like the Pareto principle, a relatively small percentage of individuals are responsible for a disproportionately large share of the impactful output or possess the highest levels of sought-after skills. The difference between an average performer and a top performer isn't linear; it's often exponential.

Does this stark reality of unequal distribution, where a select few capture the vast majority of positive outcomes, sound familiar? It strongly echoes dynamics unearthed within the world of online dating apps. Researchers analyzing user activity on platforms like Tinder discovered a pronounced power law governing user interactions, especially concerning matches received by male users. Studies, such as the one detailed in "Analyzing Massive Data Sets of Online User Activity," provided striking data: the average match rate for men hovered around a mere 0.6%, while for women it was significantly higher at approximately 10.5%. Further illustrating the skew, data showed a large percentage of women achieving very high match rates (e.g., 30% experiencing rates between 81-90%), while comparable success was rare for men. OkCupid data similarly showed women rating only about 1 in 6 male profiles as "above average."

This phenomenon gained wider cultural awareness, often discussed in contexts exploring modern social dynamics, though pinning its popularization solely on a specific TV series like "Adolescence" might be complex, the underlying data starkly revealed the uneven landscape. The inequality wasn't arbitrary; it arose from a behavioral feedback loop: men, facing low match rates, tended to swipe right more broadly, while women, often overwhelmed with options, became increasingly selective, reinforcing the power law dynamic where a few highly desired profiles attract most of the attention.

This very dynamic, driven by algorithms processing preferences and behaviors leading to highly skewed outcomes, is poised to be transposed onto the job market by AI-driven hiring systems like Mercor's. As these AI tools become exceptionally proficient at identifying not just qualified candidates, but predicting those rare "10x" top performers based on vast datasets and performance simulations, employers gain the ability to be hyper-selective on a global scale. They can focus exclusively on the candidates AI flags as belonging to the highest performance percentiles.

Consequently, the "average" candidate, much like the average dating app user, may face a dramatically more competitive and potentially demoralizing landscape – a "job market hell."

Opportunities and top compensation could become highly concentrated among an AI-identified elite, while the majority struggle for traction in a market optimized for finding the exceptional few, amplifying inequality and fundamentally changing what it means to compete for desirable knowledge work roles.

This is scary and it's coming MUCH sooner that people expect.

Data:

Key Points

• Research suggests that in online dating apps, a small percentage of users, especially men, receive a disproportionately large number of matches, following a power law distribution.
• It seems likely that this inequality arises from men swiping right more often and women being more selective, creating a feedback loop.
• The evidence leans toward significant gender differences, with men having lower match rates (around 0.6%) compared to women (around 10.5%).

The "Power Law of online dating apps" describes how matches are distributed unevenly, with a few users getting most of the attention. This is particularly noticeable for men, where a small group gets many matches, while most get few. This pattern is driven by how users interact with the apps, with men often liking more profiles and women being more selective.Gender Differences in Match RatesStudies, like one on Tinder, show men have an average match rate of about 0.6%, while women have around 10.5%. This means women are much more likely to match with someone they like, while men face tougher odds. For example, 20% of men had match rates of 6-10%, while 30% of women had match rates of 81-90%.Behavioral Feedback LoopThe disparity seems to come from a feedback loop: men, with low match rates, become less picky and swipe right on more profiles, while women, with high match rates, become even more selective. This behavior amplifies the inequality, fitting the power law pattern where a few stand out.

Detailed AnalysisThe "Power Law of online dating apps" refers to the observed phenomenon where the distribution of matches or attention on platforms like Tinder follows a power law, particularly for men, leading to significant inequality in user experiences. This section provides a comprehensive examination of the concept, supported by empirical data and behavioral insights, aiming to elucidate the dynamics at play.Conceptual FrameworkThe power law distribution is characterized by a small number of entities accounting for a large proportion of the activity, often seen in social and economic systems (e.g., the Pareto principle, or 80/20 rule). In online dating, this translates to a small percentage of users, predominantly men, receiving a disproportionately high number of matches, while the majority receive very few. This pattern is not uniform across genders, with women's match distribution showing a different, more skewed normal distribution, as noted in analyses of platforms like Tinder and OkCupid.Empirical Evidence from TinderA detailed study analyzing Tinder user activity, published in a research paper (Analyzing Massive Data Sets of Online User Activity: Who Matches with Whom and Does It Matter?), provides concrete data supporting the power law in match distribution:

This table highlights the stark disparity: men's low match rates (0.6%) contrast sharply with women's high rates (10.5%), with 86% of men's matches coming from other men, indicating higher activity from homosexual or bisexual men compared to heterosexual women. The study also notes temporal trends, such as female profiles gaining over 200 matches in the first hour, while male profiles show slow accumulation. Behavioral Drivers and Feedback LoopThe inequality is driven by differing user behaviors, creating a feedback loop that exacerbates the power law distribution:

• Men's Behavior: With low match rates, men tend to become less discerning, liking a larger proportion of profiles. Questionnaire data from the study shows 33% of men report casually liking most profiles, compared to 0% of women, and 13% of men adapt selectivity based on match rates, versus 4% of women.
• Women's Behavior: Women, with high match rates, become more selective. This selectivity is evident in their swiping patterns, where they match with most men they like, leading to a match rate of 10.5%.

This feedback loop is further supported by the study's findings on profile interactions:

• 16% of male profiles match with multiple curated accounts, compared to 6% of women, and 4% of male profiles match with over three curated accounts, indicating men's broader swiping strategy.

Impact of Profile OptimizationProfile characteristics significantly influence match rates, particularly for men, potentially mitigating some power law effects:

• Adding bios increases male matches: from 16 to 69 (a 4-fold increase) from women, and a 58% increase from men.
• Increasing profile pictures from 1 to 3 results in a 37% increase in female matches for men, and for males, matches jump from 44 to 238 in 4 hours, with female likes increasing from 14 to 65.

These optimizations suggest that while the power law is inherent, user actions can partially address the disparity, especially for men.Gender Differences in Engagement Post-MatchPost-match engagement also reflects the power law dynamics:

• Women are 3 times more engaged in messaging, with 21% sending messages compared to 7% of men.
• Median message delay is 2 minutes for men versus 38 minutes for women, and median message length is 12 characters for men versus 122 for women, indicating different levels of investment.
• Intentions differ, with 49% of men rating "one night stands" as highly important (4 or 5 on a scale), compared to 15% of women, with a statistically significant difference (KS p-value 0.00034).

Demographic ContextThe dataset analyzed included 230,000 male profiles and 250,000 female profiles, with 12% of males and 0.01% of females identified as homosexual or bisexual, focusing primarily on heterosexual users. Mean ages were 25.7 for men and 25.2 for women, with matches with 24-year-old profiles showing females matching are slightly younger on average (24.3 vs. 25.2).Comparative Insights from OkCupidWhile not explicitly mentioning the power law, an analysis of OkCupid ratings (How Do Men Rate Women on Dating Websites, Part 2) provides additional context. Men's ratings of women are symmetrical, close to a beta distribution, suggesting honesty, while women rate only 1 in 6 men as "above average," indicating a skewed distribution that aligns with the power law's implications for male attention.Implications and LimitationsThe power law in online dating highlights challenges for the majority of users, particularly men, who may struggle to get matches despite active participation. It underscores the role of platform design and user behavior in shaping these distributions. However, the data is based on studies up to 2018, and while the general principles likely hold, specific numbers and platform behaviors may have evolved by April 12, 2025. For the most current insights, further research would be necessary.This analysis integrates all relevant details from the studies, ensuring a comprehensive understanding of the power law in online dating apps, with a focus on empirical evidence and behavioral dynamics.Key Citations

• Analyzing Massive Data Sets of Online User Activity: Who Matches with Whom and Does It Matter?
• Deconstructing Tinder and the Dynamics of Online Dating
• How Do Men Rate Women on Dating Websites, Part 2