Comment by myng111
4 hours ago
This really resonates with me. I'm early in my PhD and I'm researching a niche form of data compression and IC design. I don't use any AI at all in my research, I do it the super old fashioned way, I read papers cover to cover and sections of textbooks to familiarise myself with the field.
I genuinely enjoy doing this, it's really fun to think critically about what an author wrote or how a particular approach works.
But it does make you wonder, why bother? Probably a frontier model could one shot my algorithm in a day or less. It's incredibly depressing. At least I'm not forced to use it now, but I fear I will have no choice after I join academia or industry in the future.
I can say with some certainty that you should most definitely bother.
The greatest predictor of AI productivity I have seen has been operator maturity.
AI usage tends to atrophy skills, unless used in a very mature and self aware manner.
All of which depend entirely on the operator having developed a mature theory of mind, and experience with living and working within their own head.