I'll do a PhD when I retire

August 2026

I'd love to do a PhD. Everything about it, from doing research, to being a teaching assistant, to going to conferences — except maybe rebuttals. It's good enough pay for comfortable living, a lot of freedom and exploration, a lot of growth, a lot of opportunities to collaborate with awesome people. Having a long-term goal, a north star that you can pursue for years.

Does it make sense to start a PhD which one would finish in 2031? I think not.

We're in the singularity*; we possibly will have AI better than entire research labs in 3 years; could mean one PhD can have incredible power, possibly working on 3–4 projects at a time and having 1 publishable paper per month; could mean we solve most of the problems and there is not much work to do.

Commitment issues? Some might see 4 years for a PhD as job security, some as opportunities lost. If one is 25 yo and has four more years before all her/his work becomes irrelevant, doing a PhD then retiring doesn't sound great... Or think you'd have to stay 20 years at the job, maybe a promotion every five years. Then there are people who change jobs every 2–3 years and rapidly advance in their careers.

What would one do in these times? First, be grateful one is not a high school student or beginning his/her bachelor =) Then grind and work the hardest you'd ever work in your life for the next 4 years, no matter what you'd do. Go somewhere where you can work hard and get rewarded for it, do something you enjoy.

Or nothing makes sense, no one knows what to do and how the world will look in 5 years.

* With singularity I don't mean we're going to have self-improving agents that will spiral into a general superintelligence overnight. I mean we're already at a point where in some fields it is superhuman, in some it is better than the average human, in some it's trash. And 10 years of progress can be made in 3.