While the above points are primarily about independent research, I want to emphasize again that upskilling is sometimes the better path depending on your career goals.
Seconded.
When I started to take alignment seriously, I wanted to do something that felt valuable, and work on something entirely original. I think that I gained a lot of value out of trying to perform at that level especially in terms of iterating on what I’d gotten wrong with actual feedback, but the truth was that I simply didn’t have the context required to know intuitively what new was, or what parts were likely to be difficult under current paradigms.
It was useful, because it was only after that that I tried to think about the hard part of the problem and converged on ontology translation (at the time), but having expected myself to contribute something original early, I was discouraged and procrastinated on doing more for a long time. Going in wanting to skill-up for a while is most likely the right mindset to have, in my opinion.
Seconded.
When I started to take alignment seriously, I wanted to do something that felt valuable, and work on something entirely original. I think that I gained a lot of value out of trying to perform at that level especially in terms of iterating on what I’d gotten wrong with actual feedback, but the truth was that I simply didn’t have the context required to know intuitively what new was, or what parts were likely to be difficult under current paradigms.
It was useful, because it was only after that that I tried to think about the hard part of the problem and converged on ontology translation (at the time), but having expected myself to contribute something original early, I was discouraged and procrastinated on doing more for a long time. Going in wanting to skill-up for a while is most likely the right mindset to have, in my opinion.