I know one guy who rushed through bachelor, master, and a PhD in mathematics in like half the regular time without ever breaking a sweat. Sometimes, you'd meet him in the cafeteria reading research papers for fun. You could mention any topic, and he'd know something related that one time he was super curious about and he spent an evening reading through Wikipedia. He's like a sponge who soaks up useful information. That made him an interesting person to be around. And it seems that it's also working exceptionally well for him career-wise.
I can only speak for myself and I find people who are genuinely curious somewhat endearing, even if they are not yet skilled at it. But my main decision factor would be how much effort someone puts in. Working hard? I'll help. Asking for help but not reading the tutorial links I sent? I won't help again.
I also believe it's an easy date tactic to ask the other person to teach one of their hobbies to you. You'll have lots of stuff to talk about, so you don't need to worry about what to say next. And that makes everything more relaxing and, hence, more enjoyable.
Explore and nurture one's motivation (the "for fun"), else the years of time/effort won't be allocated. Including topics, environment, tooling, metacognition, enjoyment, etc, etc.
Emphasize integration, making connections. As merely accumulating disconnected trivia provides less leverage for understanding and play, for engaging with and informing new observations, and for interesting storytelling. Rough-quantitative understanding can be helpful (eg Fermi questions), but utility seemingly plateaus low absent accompanying integrated conceptual understanding.
Integration seemingly requires surfing primary-ish literature (the "reading research papers"). Using paper intros, journal "in this issue" summaries-with-context, survey articles, definitive doorstop tome on topic X, etc. Because outreach and education content is very not sufficient (example: 5-yr old Q: What color is the Sun? 1st-tier astronomy grad student pervasive A: Yellow! Q: And sunlight? Common A: white. And then, not infrequently, a pause, and: That doesn't make sense, does it??? Two incompatible conceptions, perhaps first learned in K, persisting unintegrated into grad. Helped by popular intro astro college textbooks having it wrong, remaining decoupled from the science, for decades. The distance/gap between "science" education content, and science, can be very large. Despite some overlap in personnel, their goals, institutional support and infrastructure, culture and incentives, are rather divergent.
Integration can benefit from sheer quantity of material chewed on. Jim Keller (chip designer) commented his superpower was reading books, IIRC a book (or few?) a week, since he was 8 yo.
Cultivate storytelling skills (the "interesting person to be around").
Excellence is far easier if you can do playdates with excellent people. Especially in math. The delta between "I read the book and saw the lecture" and "I worked the problems" and "I did hours with someone wizzy in front of a blackboard" can be vast. And then there's people who could have a clue, but are babbling history/economics/policy nonsense to praise at DC dinners, lacking folks to push back. Being non-nutter can be hard without a good community. Witness researchers respected on their research focus, but gone nutter on some nearby-but-not-their-research-focus side interest. Or non-tech discussions on HN. ;)
I know one guy who rushed through bachelor, master, and a PhD in mathematics in like half the regular time without ever breaking a sweat. Sometimes, you'd meet him in the cafeteria reading research papers for fun. You could mention any topic, and he'd know something related that one time he was super curious about and he spent an evening reading through Wikipedia. He's like a sponge who soaks up useful information. That made him an interesting person to be around. And it seems that it's also working exceptionally well for him career-wise.