In both human behavior and AI systems, more effort can paradoxically produce worse results. (Anthropic’s recent Inverse Scaling in Test-Time Compute paper demonstrated that giving an AI model more “thinking” time can degrade its performance.) Likewise, psychology and behavioral research reveal many situations where “less is more” for humans—extra thought, choices, or analysis actually impair decision quality, satisfaction, or performance. Below are well-documented examples of this counterintuitive phenomenon, each drawing a parallel to the idea that additional effort or options can backfire.
Choice Overload and Decision Paralysis
Having abundant options might seem ideal, but too many choices can overwhelm us into indecision or poor decisions. In a famous field experiment, shoppers encountered either a display of 24 flavors of jam or a limited display of 6 flavors. More people were initially attracted to the extensive (24-jam) table, yet their follow-through was dramatically lower: only 3% of those who saw 24 options ended up buying a jam, versus 30% of those offered just 6 choices. This striking result suggests that an excess of options paralyzes decision-making, a finding that challenges the assumption that more choice is always better. Psychologist Barry Schwartz termed this the “paradox of choice,” noting that an overload of options often leads to decision paralysis, regret, and reduced satisfaction. In short, beyond a certain point, more choice = less action and joy – a direct human analog of “more compute = worse performance.”
Analysis Paralysis and Cognitive Overload
“Paralysis by analysis” describes how overthinking a decision can hinder or even prevent a good outcome. Our brains have limited cognitive bandwidth (working memory), so juggling too much information or too many criteria at once can overload us. Under high cognitive load, we may fixate on irrelevant details or become so anxious about getting it “perfect” that we can’t choose at all. Research by Timothy Wilson illustrates this vividly. In one study, college students were allowed to choose a poster to take home; one group simply picked their favorite, while another group had to explain their reasoning before choosing. Months later, those who chose impulsively were happier with their poster than those who analyzed their reasons—indeed, the “overthinkers” often ended up picking novelty posters they later regretted. The act of conscious analysis led participants to focus on superficial reasons (e.g. a funny slogan or catchy image that was easy to justify) rather than their true preferences, undermining the quality of their decision. This research, aptly summarized as “Thinking too much”, shows that excess deliberation can mislead us — paralleling how an AI’s performance might degrade if it over-complicates what should be a simple decision.
Intuitive Snap Judgments vs. Deliberation
Sometimes our first instinct outperforms careful analysis. Malcolm Gladwell’s popular book Blink highlighted the power of rapid cognition, or “thin-slicing,” where people make accurate judgments from minimal information. Laboratory research supports the merits of intuition in certain contexts. Psychologist Ap Dijksterhuis found that for complex decisions (like choosing among many-featured cars or apartments), participants who avoided overthinking—by distracting themselves before deciding—ended up more satisfied and made choices that independent experts rated higher, compared to participants who analyzed the options deliberately. This “deliberation-without-attention” effect suggests that our unconscious, intuitive mind can integrate complex information more effectively when we don’t try to force it. Similarly, studies on “thin-slice” judgments show that people can make surprisingly accurate assessments with very limited data. For example, observers who watched just 30 seconds of silent teaching clips could predict teachers’ end-of-semester effectiveness ratings with high accuracy. Even a 6-second glimpse of a teacher often sufficed to gauge their competence. These findings reflect the wisdom of quick judgments: in many real-world situations (from choosing mates to evaluating professionals), excess deliberation or information can introduce noise, whereas gut feelings capture the signal. It’s a human echo of the AI result that “longer reasoning ≠ better reasoning”.
Satisficing (“Good Enough”) vs. Maximizing
When deciding, one can either satisfice – choose the first option that meets the criteria – or maximize by exhaustively comparing every possibility to find the “best.” Intuitively, maximization sounds superior, but research shows that chasing the absolute best often backfires. Psychologist Barry Schwartz and colleagues found that habitual maximizers (people who relentlessly seek the optimal result) tend to end up less happy with their choices and more stressed than satisficers. In surveys and experiments:
- Maximizers reported lower happiness and life satisfaction, and higher regret and depression, even when they objectively made “better” choices. They also felt less satisfied with consumer decisions and were prone to second-guessing and social comparison.
- Satisficers, who settle for good-enough options, generally avoid this overanalysis trap and feel more content with their decisions. They spend less time agonizing yet often achieve outcomes that satisfy their needs without the downside of regret.
The more-is-better mentality of maximizers actually undermines well-being. In other words, seeking perfection through extra effort can lead to worse subjective outcomes. This mirrors the “inverse scaling” idea: beyond a point, additional optimizing (like extra compute or extra deliberation) yields diminishing or negative returns.
Flow States and the Perils of Overthinking in Performance
Elite athletes, musicians, and other performers often describe being “in the zone” – a flow state of automatic, effortless action. In these moments, they don’t consciously analyze every move; they let their well-trained subconscious run the show. Research on choking under pressure reveals that trying too hard or monitoring oneself too closely can sabotage performance. Psychologist Sian Beilock finds that when people get anxious about doing well (a big tournament, a high-stakes exam) they tend to start overthinking tasks that normally feel automatic, a phenomenon nicknamed “paralysis by analysis.” Even highly skilled individuals can blunder when they revert to conscious control of an expert skill. For instance, a veteran golfer who normally sinks putts by feel may miss if he starts scrutinizing his form in detail. Beilock’s studies show that focusing on each micro-step of a well-practiced routine overloads the brain with irrelevant monitoring, disrupting the smooth execution of skills. One striking finding is that giving experts a mild distraction (like having them hum or listen to music) can actually improve performance under pressure, because it prevents overthinking and lets implicit memory take over. In short, less conscious effort = better performance for tasks we have already mastered. This aligns with the general theme: whether in a neural network or the human brain, piling on extra “thinking” can derail processes that otherwise work best instinctively.
Quick Decisions Under Constraints
In high-pressure, time-critical situations, rapid decisions guided by experience often outperform slow analysis. Emergency responders and military tacticians know that overdeliberation can be deadly when seconds count. Research in naturalistic decision-making (e.g. Gary Klein’s work) documents experts like firefighters, pilots, or ICU nurses making life-saving calls based on gut instinct and pattern recognition. These seasoned experts don’t compare dozens of options or crunch every variable—instead, they intuitively recognize familiar patterns and act swiftly. While not infallible, such snap judgments in constrained settings frequently prove correct, because they draw on tacit knowledge that would be hard to articulate in the moment. The key insight is that past a certain point, analytical reasoning under duress can bog one down, whereas trusting trained intuition yields better outcomes. This is a human parallel to cases where “thinking longer makes the model dumber.” When time, information, or cognitive bandwidth is limited, simpler, faster decision strategies can trump exhaustive analysis.
In summary, across many domains of human behavior, adding more thought, choices, or computational steps can reach a tipping point where it undermines performance and satisfaction. From the shopper frozen by too many product options, to the student who second-guesses themselves into a poor choice, to the athlete who chokes by over-analytical self-focus, the pattern is consistent: sometimes, less truly is more. These examples underscore a deep parallel with the inverse scaling phenomenon in AI – whether it’s neurons or transistors, there are situations where exerting more effort or processing leads to worse results, and knowing when to stop is part of the art of optimal decision-making.