And special thanks to my friend @Joonsik Moon—the fun journey began with a random question over lunch with him.

📌 Trilogy blueprint
1: How does astigmatism distort orientation? 👉 buly.kr/2qZ8aCL
2: Do people with chronic astigmatism perceive orientation better than optics predict? 👉 buly.kr/15PYm2m
3: Does the early visual cortex adapt to long-term astigmatism? 👉 buly.kr/9BWPtYW

8/ 📌 Summary:
1. Observational resource is limited, so we must prioritize what we observe.
2. After a negative social experience, we allocate more resources to others, increasing negative biases over the others' intentions.
3. Where we look shapes how we understand others

7/ 🔍 We often assume we see everything during social interactions—or that more observation means better understanding. But this study shows the opposite: what we choose to observe—especially after negative events—biases our understanding of others.

6/ 🧠 So does observation cause inferential bias? Using a task-optimized RNN, we perturbed which character it observed (self vs opponent). The RNN's inferences were biased depending on observation, revealing a strong link between selective observation and social inference.

5/ Why hysteresis? Because people observed the opponent more during betrayal👁️. Eye-tracking revealed they shifted their gaze to the opponent—at the expense of controlling their own character.

4/ 😮 Surprisingly, people overestimated the computer's competitiveness only when betrayal accumulated over time. They didn't exhibit a similar bias toward cooperativeness—even when the computer consistently assisted. This path-dependent process is often called hysteresis 🔁

3/ Participants had to guess the computer's intention, then decide whether to speed it up or slow it down. Unexpectedly, the computer's intention changed after their decision, resulting in betrayal 💔 or unexpected help 💡.

2/ We built a naturalistic, socially interactive Pac-Man game 🎮 (inspired by Heider & Simmel, 1944, and Yoo et al., 2020—
@myoo.bsky.social, @benhayden.bsky.social ), where people hunted prey with a computer opponent who could be either cooperative 🤝 or competitive 🥷.

1/ Our paper asks: How does our choice of what to observe 👀 in social situations shape our understanding of others' intentions? We find that selective observation can lead to lasting inferential biases—especially negative ones.