There are two kinds of people that I know:

Lab Tinkerers

Problem-Solvers

These two kinds of people are extremely similar; it may even seem like the same at first. But I want to make a clear distinction as somewho has tried both math problems and lab tinkerers. To understand, let's break down each role.

Lab Tinkerers: You spend your time making tweaks to an idea to bash your head on the wall when the performance doesn't improve. Maybe a threshold value of 0.8 will work better than 0.6. Oh, the testing still shows statistical insignificance. You do contests like ISEF or JSHS, submitting your tinkering to the project to showcase how you got some cool results.

Problem-Solvers: You spend your time making ideas from an idea, bashing your head on the wall when you realize it doesn't work or solve anything. Maybe doing this trick will lead to a functional answer... Oh, you have no idea where to go from here. You do contests like AMCs, USACO, USAPhO, USNCO, trying to find problems to make you scratch your head.

What is the key difference? The answer is simple. One is looking for a solution, while the other is attempting to satisfy a metric.

Where was I?

In high school, I really wanted to love problem solving. But the matter of fact was that it just never clicked with me. I didn't really have a lot of support on this matter since my dad (Chemistry) was on the lab-side of work. So yeah, overall, I just didn't really like bashing on a piece of paper. I really wished I qualified for AIME or got higher like USAMO. Or maybe something like USACO Platinum or IOI. Well, I didn't.

But if there's one thing I did like, I really enjoyed tinkering. I would stay up till 1:00, messing around with AI parameters to try to get a good accuracy. I loved to make tweaks to code, trying to get some tests to pass. I never did contests like ISEF or JSHS (more on this later), but at the same time, I got a really good amount of value from high school, landing me an interview @Epic (healthcare), Underwriters (insurance), and a few other companies; pre-college as well as some publications and a ongoing paper to NAACL SRW + 2 research internships for strong placement in college.

Let's talk about future prospects of a realistic comparison between problem solving vs tinkerer.

Problem Solver

• Top 5 Olympiad Summer Program [Math, CS, Chem, Phys, Bio]
• RSI + ISEF in STEM
• Graduate Classes in many STEM
• Goes to MIT/Harvard

Tinkerer

• 2x SDE interns at good company as a HS student, 2x Research Internship at University
• Paper published in SRW conference of a top field (SIGCOMM, NeuralIPS, etc) as first author
• Graduate Classes in a single STEM on PhD track
• Goes to Stanford/Yale

Career Paths (a possible one):

PS1: hft, hft, hft, hft -> FT HFT at probably >500K/yr.

TK: Google Research, OpenAI, Deepmind, Deepmind -> FT research at deepmind at probably >500K/yr

Takeaway: Career scaling with either is not going to be a problem.

Which one of these is a harder track? It depends on the person. There is 2 bullet points that are extremely hard to get (IMO + ISEF Grand Award VS Paper at main conference and 2x SDE at top company + 2x Research.

For me, the tinkerer track is significantly more realistic. Since I was born, I was disqualified from qualifying for the IMO. 

Going back on topic, let's talk about Lab Tinkerers vs Problem Solvers.

The great thing about computer science is that it is a field that allows both of these kinds of people to exist. If you go to a highly theoretical field such as computation research or distributed algorithms, you end up going down the problem-solving route, where the only thing you can do is bash your head on a wall. If you go down a tinkerer route like systems, you spend your time making tweaks to the algorithm code to bash your head on the wall.